GPLv3

Comment 3151: Bracketed section numbers

Please remove these for the final version of the license.

Comment 3152: Ambiguity

'it' is ambiguous, does it refer to the server or to the corresponding source?

Perhaps "You remain obligated to ensure that the Corresponding Source is available for as long as needed to satisfy these requirements, regardless of what server hosts it."

Comment 3153: Joke isn't funny, and is offensive to some. Remove or change this name.

Some find the term "Coon" offensive. Having lived my entire life with a surname that is a homonym for this racial slur, I can tell you with strong anecdotal evidence that the word itself does make many people, particularly those who grew in areas of the USA where the slur was commonly used.

I can't easily change my own surname, so I live with it, but there is no reason to use this word unnecessarily, particularly when it is even spelled the same as the offensive racial slur.

I'm all for making people uncomfortable when pursing a greater good -- like raising awareness for a political cause or a moral issue. However, this is offending people just to make a (frankly very) lame joke. Why bother? Why not simply take it out?

This has been asked for several times; I suspect it isn't going to happen until we can come up with a suitable replacement which tickle's Richard's funnybone.

The closest I can get is: Ty Rant

Also: Moe Gul? N. Trepreneur? C. Zar? X. Ecutive? N. Vestor? Lee Der? O. Verlord? O. Verseer?

Or, given that it's been a man for over 15 years... Di Rector?

I think this is the ultimate [dis]proof of the claims that the comments process has been meaningful. That they have refused to change this trivial-to-the-license and obviously offensive text is just appalling. The only possible explanation is that he-who-must-be-obeyed thinks it's funny...

Please change the name.

Aye, Coon is extremely un-funny and extremely offensive. Having been directly called that by members of the KKK as they burned crosses on my lawn and threaten to lynch members of my family, I have zero love for the word.

"Mr. Smythe" is a great alternative for this.

I frankly DON'T find the words "Ty Coon" offensive, in the harmless context, in which it is used, in the GPL.

The two words are meant to be a play on the word "Tycoon", which, according to Dictionary.com, means "a businessperson of great wealth and power; magnate." It is obviously meant as a joke about a tycoon with a similar sounding name. I don't see how anyone could possibly reasonably be "offended" by this.

Listen, this has been in the GPL for at least 16 years and no one has ever had a problem with it. In my humble opinion, there is no need to change it - leave it as it is.

Comment 3154: Subsections required?

Would clarity be improved by having '7.1 Additional Permissions' & '7.2 Additional Terms'

I agree; the two halves of this section are less closely related than other areas of the license where two sections are used instead of one.

I agree, but s/Terms/Requirements/, as Terms is too neutral. This section is *extremely* confusing to the new reader. I had to go back to early drafts and rationales to understand it, as the current draft muddles permissions and requirements, which are treated entirely differently.

I think the clauses should be put in the relevant places in the license, not all lumped together simply because they are additional terms. See my Comment 3445: Restructure the GPL in terms of Rights and the associated Responsibilities

Comment 3155: Change of address issues

We had planned to get a PO Box for use in the GPL that could stay the same in case FSF needs to move its offices again. Well into the 2000s, people were still sending postal mail to the Mass Ave. address (which was vacated in the mid-1990s). Indeed, google still shows that address associated with FSF all over the Internet, and Temple place is even more prevalent. The new office space is wonderful, and I hope FSF keeps it for a long time, but the lease is (ISTR) only 5 years, and GPLv4 might not happen before the lease is up. And, GPLv3 may even stay in common use for years after GPLv4 is released.

Why not get a PO Box in the main post office in Boston or Cambridge for the posterity of GPL? Surely the donors to FSF won't mind paying this small annual fee.

This is equivalent to comment 2641 on Draft 3.

I agreed with this comment on Draft 3; I agree with it now again. I, too, will re-raise issues I had with Draft 3 that are not solved in this Draft. There were many pertinent, even important, comments on Draft 3 that are not addressed.

Can the FSF please publish a rationale, not on the changes, but on the reasons why many thinks did not change, despite the comments?

Comment 3156: please remove the word Coon

please remove the word Coon

Comment 3157: Needs to be "User Product"

Let's imagine I am making a User Product whose primary purpose is to communicate over a network. There is a networking chip with proprietary firmware in ROM which manages the voltages on the wires. The product's main processor runs a GPLv3 OS and kernel. The kernel has a device driver which exchanges packet information with that chip.

I can write a proprietary boot loader which checksums my kernel and, if the checksum fails (i.e. the kernel has been modified), sends a signal to the networking hardware to stop putting packets on to the wire, thereby rendering the product useless. The chip continues to behave normally in all other ways. This is not forbidden under this wording. The object code is functioning exactly as it did before - it is passing networking packets to the networking chip. So this system does not fall foul of this clause. I think this is a large loophole, and permits Tivoisation schemes. I believe changing these two words to "User Product" will fix it. The whole product must continue working when the code is modified.

Comment 3158: Word used before it is defined

The intermingled definitions of Standard Interface, Major Component and System Libraries are confusing, perhaps partly because Major Component is (needlessly) used before it is defined. These two paragraphs need expert attention from a wordsmith. I am not asking for any changes to the meaning, just improvements to the English.

Comment 3159: Section titles for 4, 5 and 6 need to be made consistent

Sections 4 and 5 both apply to source code, and section 6 to object code. However, the title of 4 does not say so (whereas 5 and 6 do) and the title of 6 does not use the pre-defined term "object code". I suggest the following set of consistent and progressive titles:

4] Conveying Verbatim Source Code.

5] Conveying Modified Source Code.

6] Conveying Object Code.

I agree that more consistency is needed (though not necessarily as suggested here). Further to the inconsistencies mentioned here, the terms used in these section headings, and those used within the body of Section 2 (which speaks of "the unmodified Program" and "covered works"), are also inconsistent.

Comment 3160: "extensions of a covered work" is not defined

This phrase is not defined in the rest of the license. What does it mean to be an extension of a covered work? There are many possible interpretations. Does it mean "will not run without the presence of the covered work", "is designed primarily to use the covered work", "can run only on the output of the covered work", or something else?

I have the feeling that they're trying to be vague so that they can claim as broadly as they like...

I agree this is too confusing. Just make it "A compilation of a covered work with other separate and independent works, which are not licensed under this license..."

Comment 3161: Series comma needed

Change "run and" to "run, and"

Ummm, they are correct in the way they have it written... You do not put a comma after the word "run"....

There are two ways to write series of terms: "A, B, and C" or "A, B and C". Careful writers tend to prefer the former (sometimes referred to as the series comma or the Oxford comma) because it avoids the ambiguity about whether the final term is a combined term (does "cabs come in yellow, orange, black and white" mean that some cabs are "black and white" or that some are black and some are white?).

This document uses the series comma pretty consistently, but the changes in this revision introduced a couple of places where it was left out, of which this is one. Hence the request to make the punctuation consistent with their style elsewhere. "Run" and "propagate" are independent, so they should be separated with the comma.

Comment 3162: Very good

I'm very glad to see this clause!

Comment 3163: Extend to collaborators as well as contractors

It seems to me [IANAL] that if collaborators (say Fred and George) are working together on developing a new release of a covered work that includes material copyrighted by another party (say Jim), that the requirements of sections 5 and/or 6 would be triggered each time they passed incremental work back-and-forth between them. This seems pointless.

For instance, every time George sends Fred an executable to test, George would have to archive the current state of the source in case someone else should happen to obtain a copy of the executable and request the correspondinf source.

Perhaps you could add "or for the purpos of making modifications on which you will jointly hold copyright" after "exclusively for you" and modifying the following sentence to read: "Those thus making or running the covered works must do so strictly in conjunction as part of that development or operation, on terms that prohibit making any copies of the material outside their mutual relationship."

A better way (IMHO) would be to make a special exception to slightly relax the requirement to convey source, and allow a later version of the source to be conveyed instead, if the object code was originally conveyed as part of the development process, and marked as a developmental version.

Comment 3164: Prefer "that" for restrictive clauses

I know it's pedantic, but using "that" makes it clearer that this is restrictive rather than descriptive.

My bad - I highlighted the wrong "which" when I wrote the comment - it's the second one ("which is normally") that should be "that" ("that is normally").

Comment 3165: Right of determination of violation

The language in the preceding sentence and this sentence are problematic. THe preceding sentence says that "giving notice" is predicated on a violation, but, in fact, it would be conditioned on the copyright holder believing such a violation occurred. This sentence then says the owner may terminate the license some time after notification, but, in fact, such termination would presumably require also that there be an actual violation.

This sentence arguably suggests that the copyright owner could terminate the license even if there were no real violation, based simply on having given notice.

I agree with sepreece. I think that it should be clarified.

sepreece is totally correct. When a copyright owner can terminate the license to individuals arbitrarily, even in the case where there actually is no violation, this is in itself against the very tenets of the GPL and Open Source software as a whole. Programmer "A" hates User "B". User "B" is using Programmer "A"'s program. Programmer "A" says "User B has no license anymore - I hate him". It should be specifically clarified that the ONLY time a copyright holder can terminate the license is in the case that there is a known, proved, documented and investigated violation. The paper trail of the documentation and investigation of the claim of violation should be necessary.

It seems to me that having been "put on notice" as defined here, does require an actual violation, so the paragraph works, even if it's a bit weird. (If you receive a notice, but it turns out you aren't actually in violation, then you were never "put on notice" in the GPL sense.)

/ When a copyright owner can terminate the license to individuals arbitrarily, even in the case where there actually is no violation, this is in itself against the very tenets of the GPL and Open Source software as a whole. /

Given the complexity and convolution of the license, the sheer number of requirements and additional possible requirements, the ambiguity and vagueness of some of these requirements, and the fact that people are not infalliable, I think the right to terminate the license when there /is/ a violation is only slightly less problematic.

Comment 3166: Need a comma here

I think you need a comma here (after "of") to make the pieces of the sentence play together nicely - the three clauses all flow into the "one or more...", but without the comma it appears that only the final one does.

Comment 3167: series comma needed

Should be "modify, or convey"

Comment 3168: Glad to see this

Glad to see this finally nailed down.

I disagree because this section will prohibit indirectly the users from moving to the next version of GPL due to fear of losing of patent protection from the copyright holder

Isn't a comittment to honour a later version of the license itself an obligation? I think this needs to expicitly allow further obligations of "negative liberty"--i.e. obligations not to interfere with others' liberty (as opposed to "positive liberty", which are obligations to proactively uphold others' liberty)--in order to allow license upgrading to actually be effective.

Comment 3169: Fix wording

"Responsibilities" is inconsistent with only mentioning a single responsibility ("to respect..."). Change "to respect" to "in respect of" or simplfy the sentence to "Therefore,you have a responsibility, if you distribute copies of the software or modify it, to respect the freedom of others".

How about changing it to "responsibilities which respect the freedom of others".

There are multiple responsibilities, including allowing bypass of DRM, allowing use of patents, and in some cases, providing source. I think it's ok as is.

Comment 3170: Series comma needed

Should be "industrial, or non-consumer"

The word "non-consumer" should be eliminated as this word is unnecessary and relies on the literal definition of the word "consumer".

Comment 3171: soften the prohibition

Instead of requiring the terms to establish such a prohibition, how about merely *permitting* such a prohibition in those terms?

Because if they were not prohibited from conveying the non-GPL'd segments of the work, they could conceivably distribute those segments under terms differing from the GPLv3, I assume.

Comment 3172: include in its coverage the non-exercise of rights?

It took me a while to connect the "coverage" with "one or more rights" rather than "non-exercise". Maybe I'm just slow, but if there's a way to clarify this sentence...

If they put the clauses that end with dangling "of" and "on" first, and the one ending in "coverage" last, that would help, because they much more obviously are waiting for completion. Adding a comma after "conditioned on" would also help, as noted previously.

Comment 3173: as I see it, this is a loophole

As I see it, a loophole can be found here.

If the third party is not in the business of distributing software, but holds a patent portfolio, a discriminatory license is still possible. Such a party could also be used as intermediary.

This whole discriminatory patent license section is too specific in my opinion, and individuals looking for weak points can use that to their advantage.

I agree that the marked clause should be dropped, but I think the section is otherwise good.

Comment 3174: man-in-the-middle third party

Does this paragraph stop a Free Software enemy who holds patents and a Free Software distributor from finding a third party willing to serve as a man in the middle?

E.g., a company that isn't in the business of distributing software and that doesn't hold or control any patents, but that enters agreements with both Microsoft and Novell such that Novell's customers get Microsoft's patent licenses and funnels the corresponding payments back and forth?

Usually anti-racketeering laws prohibit these types of business relationships from being formed, and are normally illegal in most countries in the world. However, we know that companies with sufficient funds to buy or bribe whole governments, can circumvent these restrictions. A specific clarification on this issue should probably be done within this paragraph to specifically eliminate this loophole, and any others that could be thought of. Some general wording terminology could suffice - or maybe it does need to be very specific. Anyone with ideas, please chime in.

Comment 3175: What is a user?

Please define 'user'.

A "user" is an individual who owns a legitimate copy of the software to which this licence is affixed.

A "user" is not necessarily a licensee, however, only users may become licensees.

A "user" does not need to actually use the software in order to become a licensee. Similarly, mere use of the software does not define or confer ownership of a legitimate copy.

A legitimate copy is one that has been obtained with the consent of all its conveyors including the licensor and user from which it was obtained.

The idea of a "legitimate copy" proposed here is no good. We should not have to trace backwards through all the conveying that led to a specific copy, and determine that it was all legitimate, in order to determine that the copy in question is legitimate. It is enough that the work was licensed by the copyright holders. What if I copied OpenOffice from a second-hand laptop which I didn't realise was stolen? This would mean I now have to recopy (an identical copy of) OpenOffice in order to have a legitimate copy. That's got nothing to do with the real issue, which would be the return of the laptop.

Fair point.

So, add:

A legitimate copy is also one that has been obtained with the consent of its immediate conveyor where the copy is identical to a copy that is, or has been, available to members of the public.

Buy a stolen laptop containing unpublished GPL derivative - the copy is not legitimate.

Buy a stolen laptop containing published GPL derivative - the copy is legitimate.

Comment 3176: Too soon

Having this clause show up so soon sends the wrong impression and leads to confusion.

Comment 3177: Much better title

This is a much better title, given the goals of the section.

Comment 3178: Here be dragons

Copyright itself originally envisaged application only to a few wealthy and industrious printing press owners. 300 years later copyright applies to everyone. Something tells me they might have benefitted from hindsight.

You really shouldn't start splitting hairs when it comes to deciding the difference between 'consumption' and 'commerce'. We are all industrious, commercially oriented beings. Don't denigrate the individual by imitating those who would treat people as passive 'consumers'. Art is not consumed, but lasts forever.

Today and ever more frequently we will start seeing software products designed for collaborative use by collectives, e.g. Wikis, etc.

A Wiki is normally used for community or collaborative purposes. Are you suggesting that a Wiki is exempt because it isn't used for personal, family or household purposes?

I am dismayed that the GPL now considers liberty more deserved by the cash rich materialist 'consumer', than cash poor hobbyist software engineers selflessly working for their communities.

Not all non-'consumer' uses are by wealthy corporations, there are many poorly funded communities using software products (for which no significant personal use occurs) awaiting exploitation of this loophole.

As I suggested before, if a corporation wishes to surrender their liberty then they can rent their IT facility.

Let's keep the GPL simple and straightforward - if you buy a software based product, you own it and the software within it, and are entitled to the same liberties whether individual or corporation, and whether for domestic or commercial/industrial use.

There is no such thing as a consumer of software or software based products (unless we're talking about endoscopic pills http://www.medgear.org/entry/robot-that-can-be-swallowed-to-test-for-cancer/ ). The word 'consumer' should not appear anywhere in the GPL. It is intrinsically insulting in this context.

Of course, if the FSF is receiving significant funding from a wealthy corporation with a considerable client base who require this discrimination between consumers and corporations, well, ignore everything I've just said. Let's keep things in perspective eh? I could do with a new Ferrari myself.

I agree with this sentiment. I believe this clause exists because there are machines that must remain unmodified for safety reasons, e.g., medical devices.

My thought is those medical devices should clearly indicate to users that it has been tampered with, instead of ceasing to work at all. Manufacturers can use this to their benefit by having a green light that says the device passed internal safety tests, vs. a red light saying it has not. This preserves the necessary tampering protection without removing any user rights.

Even if we ignore the fact that people who work in industry have rights (and responsibilities) just the same as people who do not, any device that is purely commercial today might be taken home for personal use later.

For example, someone might buy an obsolete MRI machine and modify it to better scan their cat. Certainly this may make it less accurate for humans, and the device should clearly notify users that such is the case. But the cat lover does own the device, and his freedoms would be harmed if he could not modify it, not to mention the harm the lack of these modifications could cause to his poor, sick cat.

If the device instead fails to work when tampered with, this creates an incentive to disable the anti-tamper hardware so the device can be modified for use on cats. This results in an unsafe-for-humans device that does not notify its users in any way. If the cat owner later dies, and the device is then sold to a 3rd world country for use on their citizens, this GPLv3 clause (when compared to the alternative) could ultimately cause great harm.

I agree that "consumer" implies a selfish view of what the buyer is, and that is why "consumer" is not taken literally in this definition so it applies properly regardless of one's definition of "consumer". "Are you suggesting that a Wiki is exempt because it isn't used for personal, family or household purposes?" No, this part of the license does not target Wikis.

yuhong, I gave Wiki as an example of a non-'consumer' software based application or device.

Who's to know what sort of collaborative devices could be created in the future?

Someone could produce a collaborative mapping device - that had absolutely no utility except when used collaboratively.

One could imagine a community version of something like Google's StreetView, and a manufacturer producing a GPL based photogrammetric survey camera. Now because this community can demonstrate no personal/domestic use, the manufacturer can say "Tough. Even if you could afford to buy one, you can't modify the GPL software in our device. It's industrial, so there."

I am objecting to this subjective and spurious distinction between classes of users, when there should be no distinction at all. If you have a copy, you have a licence. If you sell a copy (or a device containing a copy) then you must observe the terms of the licence. Rich or poor. Industrial magnate, corporate professional, community member, or keen amateur.

The last thing the GPL should be doing is bending over backwards to accommodate those who don't want some of the liberty the GPL restores to them. For such people who don't want the rights and responsibilities of ownership, the solution since time immemorial is RENTAL!

Comment 3179: Prevent loopholes

I would be more comfortable if this sentence also contained something to the effect of "and any other activity permitted by this license".

the problem herer is that the covered issues are not part of the exclusive right which is the patent.

Also specify that the right is to do these things under the terms of the license -- it is not a grant to use patents in non-GPL works.

Comment 3180: Belongs in an FAQ

The license text has gotten substantially longer than GPLv2. Also, since 1991, the Free Software community has grown such that it is either common knowledge how to do this, or easily found in FAQs or other documents. There is little point anymore to carry this section with the text of the license officially.

"There is little point anymore to carry this section with the text of the license officially." I disagree. What if someone doesn't have internet access to view an faq or a document? We shouldn't assume that someone knows how to do this, or has resources like the internet, etc... to find out. I think that it should stay.

I agree with splscool that just because some people have ready access to, or have used the GPL before in the past, new people are always coming on board and using the GPL for their programs and projects, and having the "official" directions on how to do it all in one place, rather than having to search around through unrelated information in order to find it is a good thing. Clarity is good.

I pointed someone new to Free software at this section just the other day. Yes, perhaps I could have pointed them at a FAQ. But having it hear means that the headers on lots of GPL software are pretty consistent. And that's useful.

Even though I don't think this section is entirely necessary, if it is to stay then it should be cleaned up....

Comment 3181: is this a blanket release?

Perhaps it was intended to write: "...them do so under that particular interactive interface." rather than just "...them do so."

Comment 3182: Should this be against and plural?

Offhand, 'Protecting Users' Legal Rights Against Anti-Circumvention Laws' reads better to me than 'Protecting Users' Legal Rights From Anti-Circumvention Law.'

Comment 3183: if verbatim is too confusing here, it is probably too confusing in s. 4

I assume that the intent of this was to reduce the use of the overly-lawyerly 'verbatim', but given the critical use of 'verbatim' in s. 4, maybe it is better to be consistent across the sections and use either 'exact copy' in both places, or verbatim in both places?

Verbatim would be better, in part because it better carries the notion of "same text", where "exact" might suggest that you would need to provide the same physical packaging. I.e., a "verbatim" copy of a book would be a copy of the text, an "exact" copy of a book would have to match the typography, cover art, etc.

Comment 3184: Letting Novell Off the Hook

Allowing Novell to continue with their Microsoft deal will most likely prove be to the detriment of GNU/Linux. It simply makes no sense and is unjust to the GNU/Linux world to allow them to get away with making such a blatantly negative move as the Microsoft deal is against the GNU/Linux community. They certainly will not hesitat to crush free software in the future if they get the chance and think it will raise their bottom line. This escape clause for Novell makes no sense.

I do not like how this date embedded in the license gives a feel that it was randomly picked and without any further explanation it gives the impression you're trying to "sneak" this in here.

There are a lot of patent license aggreements that need to be sorted out for open source. Putting this clause is just leaves problems.

I think that this line is good just because there ARE so many outstanding agreements out there. This allows those companies to continue business with what they previously have created, but blocks any further monkey business from ever happening again. Think of this as your generic "grandfathering clause".

Personally I dont like giving this out to Novell, but Novell is not the only company affected by this.

As can now be seen in the News: http://news.google.com/news?hl=en&ned=us&q=grandfather+clause+novell&btnG=Search+News

The vultures are starting to circle.

There is no telling what kind of bad place this grandfather clause for Novell is going to take Free Software to.

Novell, LG, Xandros, Linspire. I think we either take a firm stand for software freedom against these practices, or succumb to being divided and defeated, making GPL a joke.

Please remove this exception. The reasoning that these deals will eventually backfire against Microsoft is dubious and full of risk.

/ Novell, LG, Xandros, Linspire. I think we either take a firm stand for software freedom against these practices, or succumb to being divided and defeated, making GPL a joke. /

Having the date in makes absolutely no difference to LG, Xandros or Linspire. The FSF /are/ making a firm stand. They are saying "it stops here". Freedom includes the right to do anything that is not forbidden. These patent deals were not forbidden. Now they are. The FSF should not aim to be Microsoft, the FSF should aim to be /fair/. LG, Xandros, and Linspire knew the rules, and should pay the price. As for Novell, their customers can tell them what they think.

I think it is stupid to ''forgive'' Novell for making the deal and exclude it from following the GPLv3. The FSF should take a strong stand against the deal made by Novell and they will either have to follow the license and end the deal or not distribute any GPLv3 software.

Comment 3185: Website address is incomplete

This was clearly brought up with the previous draft, but still it's not fixed. Every geek knows that there must be a slash at the end i.e. . Please fix it, it's like omitting a full stop at the end of a sentence.

It's also common to use angle brackets with URLs (also mentioned with the previous draft) but that's a minor thing.

Quoting myself: at the end i.e. .

This comment system ate the URL, because I put it in angle brackets. Retry:

at the end i.e. (http://fsf.org/).

No, there is *no* requirement (nor, in general, a good reason) for a trailing slash. Look up the relevant RFCs.

I've noted this before. Although blinken says that there is apparently no reason for a trailing slash, in reality, there is. For a simple reason. By referring to the root address of your Web site, you are referring to whatever is the default page in that folder. Since we don't know what the default page is, we must refer to the directory itself, so that we can be redirected to the appropriate page. When we link to a directory, we must always include a trailing slash, for we're otherwise erroneously referring to a file with the same name as the directory.

It's a style issue, and minor, but the license should still be fixed. After all, the GPL v3 is for computer scientists, programmers, and other people who know and will care about this.

So in conclusion, I suggest changing this to (http://fsf.org/).

Comment 3187: which vs that

IMHO, this should read "..., which means any tangible personal property that is normally used...". The first "which" is definitely correct; the second should probably be replaced with "that" because it starts a restrictive clause.

Comment 3188: Contradiction with OSI definition of Open Source

Isn't this actually in contradiction with the OSI definition of Open Source at : http://www.opensource.org/docs/osd which states :

"6. No Discrimination Against Fields of Endeavor

The license must not restrict anyone from making use of the program in a specific field of endeavor. For example, it may not restrict the program from being used in a business, or from being used for genetic research. "

Gilles

No, it's not a contradiction. You can use GPL'd code to implement DRM. This section just says, that the GPL'd DRM implementation can be "circumvented" without violating DMCA, EUCD, or similar laws.

It would be absurd if GPL'd code could be "effective technological measure", since the source code must be available, and from that it is easy to make a "circumvention tool". If such modification was prohibited, it would render the application non-free.

I'm not sure that "making the licensed software useless for a particular function" wouldn't be regarded as "restricting its use". However, I'm not sure that it matters - I can't think why a lawyer would be arguing over whether the license was or wasn't OSI-acceptable.

My personal belief is that the clause is pointless to begin with, since the DMCA and similar laws give the copyright holder, not the DRM provider, sole authority to allow access, but I think the FSF views this point as making a statement rather than having any practical force.

Let's imagine that libdvdcss were under GPLv3-draft4. If MPAA distributed it, MPAA would no longer be able to sue anyone under DMCA if someone uses or distributes libdvdcss. (Maybe MPAA is a bad example; substitute it with "copyright holder" if it is more logical that way.)

This is a discrimination against a field of endeavour only if preventing GPlv3 code being used in proprietary software is a discrimination against a field of endeavour.

GPLv3 code may not be used in a way which restricts the users freedom. Do you call that a field of endeavour?

As larzhu says, IF the copyright holder were distributing the GPLv3 implementation, there would be an argument that this implicitly authorized circumvention. However, that seems unlikely to be the case; more likely the DRM implementation woudl be in a player, not distirbuted by the copyright holder, and no such implication would exist.

Larzhu's comment totally misses the point. And his example covers only a rare case, probably totally irrelevant. What the license says is that the programmer recognizes that his software is there only to help circumvent whatever protection measures is used on the files accessed by the work. Hence he is liable for [contributory] infringement of copyright laws as soon as the software is released, which, in my country, an d probably other, will fetch up to $360 000 and 3 years in jail. Conclusion : any sane programmer will avoid ever using the GPL v3 to be sure he never is blocked by the threat. And contrary to Larzhu's naive belief, the status of effectiveness of a technical measure is legal and not technical (there is no such thing as a technically effective measure). It does not exclude the accessibility of source code. The fact that the license cannot prohibit any modification of the program does not mean that any modification is legal. Some programs may not be legal ... no matter how obtained. Larzhu's last argument dead. BTW, why bother having a clause that prohibits implementing effective technical measures, if it were so obvious it cannot be done. So Larzhu's arguments are not only wrong, but also inconsistent.

BTW, is it OK under GPLv3 to use the work to manage the security system of a jail and restrict the freedon of people ? Maybe the inmates are not considered users.

Larhzu's comment totally misses the point. And his example covers only a rare case, probably totally irrelevant. What the license says is that the programmer recognizes that his software is there only to help circumvent whatever protection measures is used on the files accessed by the work. Hence he is liable for [contributory] infringement of copyright laws as soon as the software is released, which, in my country, an d probably other, will fetch up to $360 000 and 3 years in jail. Conclusion : any sane programmer will avoid ever using the GPL v3 to be sure he never is blocked by the threat. And contrary to Larhzu's naive belief, the status of effectiveness of a technical measure is legal and not technical (there is no such thing as a technically effective measure). It does not exclude the accessibility of source code. The fact that the license cannot prohibit any modification of the program does not mean that any modification is legal. Some programs may not be legal ... no matter how obtained. Larhzu's last argument dead. BTW, why bother having a clause that prohibits implementing effective technical measures, if it were so obvious it cannot be done. So Larhzu's arguments are not only wrong, but also inconsistent.

BTW, is it OK under GPLv3 to use the work to manage the security system of a jail and restrict the freedon of people ? Maybe the inmates are not considered users.

I bite. While I don't haven't studied law, I'm pretty sure that I understand this part of the license correctly.

First, if you write a GPLv3'd DRM cracking tool, it *is* a circumvention tool in terms of law. However, if you convey such work, you waive your right to sue *anyone* under the appropriate part of DMCA/EUCD/whatever for using the work. As sepreece already wrote, this usually has little practical effect, but I cannot see anything bad having this clause there anyway.

Second, if you create a DRM implementation under GPLv3, you cannot sue anyone if they circumvent it. That's because when you convey your work, you have waived your right to sue under the appropriate part of DMCA/EUCD/whatever. Thus, if someone uses GPLv3'd DRM implementation, anyone else is free to circumvent it. I cannot see anything bad in this (unless you like DRM).

Quoting lang: "What the license says is that the programmer recognizes that his software is there only to help circumvent whatever protection measures is used on the files accessed by the work."

No, you have misread it badly. I hope you don't think that liability disclaimer in in the section 16 means, that author recognizes that his software is there only to harm people's computers.

Comment 3189: Terminology used is no longer accurate

The old terminology "general-purpose computers" is no longer accurate nor viable in today's evolving society and technical world. The terminology of the "general-purpose computer" is traditionally aimed at large scale system level computers and mainframe computers, although some looser interpretations allow the extending of that definition to the class of computers termed "workstations" or "personal computers". This is a misnomer and does not work when talking about modern devices and modern software. At best you can call anything a "device that can execute software" - computational requirements are not necessary, since software itself can provide the computational requirements that may be missing from the executor. It can be a single IC chip - they are easily programmable by using a simple terminal program. It does not have to be a "computer" of any classical sense whatsoever. It can be a series of circuits that do level shifting, it can be a mechanical device that allows grains of sand to fall. This also applies to all cases where the word "computer" is used throughout the GPL.

Much of the contention over GPLv3 was about TiVoization, which specifically affects primarily devices that are NOT general-purpose computers, though they obviously contain computers. The use of "general-purpose computers" in this one paragraph of the preamble adds nothing - many of the devices that the FSF is concerned about patents applying to are not general-purpose computers - and arguably muddies the intended scope of the license. A patent infringement suit could cite that paragraph as evidence that the license's patent language was not intended to apply to special-purpose devices like phones, TiVos, routers, camcorders, etc.

I would urge dropping the phrase "on general-purpose computers" entirely.

I agree... Isn't "software" enough? <> seems to suggest that patenting software could be allowed on computers that are not "general purpose".

Comment 3190: Lack of Freedom of the license (meta-license)

Free Software needs a Free license. The FSF considers non-Free software unethical; each and every argument to this also applies to a lawyer’s text: a non-Free license is as unethical for exactly the same reasons.

Even more so since Lawrence Lessig is preaching that Law is Code (or something very similar) (with its own syntax, and logic, and side effects, etc).

The single most important issue I had with Draft 3 is not addressed, despite several comments on it: comments 2501, 2567, 2642, 2957, 2988, with several people agreeing.

Of course we do not want many forks with small different changes anymore than we want many forks of software project with only small changes. But the positive possibilities of Derivative Works of this License, borrowing either a single paragraph or an entire concept, adapted to a certain specific need, are beyond our imagination.

There is no good reason we should not allow derivative works of this License, as long as those texts do not call themselves “General Public License”.

[And if the FSF refuses to change this, I would very much like to them to explain their rationale.]

The license is not a functional work and therefore modifying it is not necessary. We need free software because we must have full control of our computers. People will confuse similar free software licenses even if they have different names. That could be serious blow to the free software community because people then will forget their freedom even more. Similar thing happened when people started calling GNU system Linux. The name Linux does not lead people to the philosophy of GNU software. Licenses that do not have GNU in their name also would not lead people to the philosophy of GNU. Therefore modifying the license should not be allowed.

"Functional work" (whatever that means) or not, why not give the community the chance to make variants of the license if they like to do so? As they say, writing GPLv3 has been a community effort too. Preventing using GPLv3 as the starting point will do nothing to prevent people from participating to license proliferation either, but at least they might get something right if they have a strong basis.

Misattribution is the real problem, which won't need a "do not modify" clause for it.

I believe it's clear that the text is referring to changing this license's text, and still naming it GPL, as opposed to using paragraphs/chunks of the license in variants not named GPL. I don't think the text should be changed in this case.

mjuarez: If the text would say what you think it says (clearly even, according to your claim), this problem would be non-existant.

But, no matter how I read the copyright statement + "changing it is not allowed" text, I can't see how this paragraph could possibly say what you think it says.

stikonas: You must be using a pretty narrow definition of "functional" if you call a license "not a functional work". Law is not art for art's sake. We need free licenses to control of the continuation of freedom as much as we need free software to have control our machines.

The main use case for this issue is (in my mind) someone (person or organisation) wanting to create a license for a type of work, which is not software (but may still be functional according to stikonas' definition) and for which a GPL-like free copyleft license is a good idea, but where not all of the GPL's software-oriented language would fit.

The GPL's copyright notice is not only hypocritical, it also means the GPL cannot legally be incorporated into GPL works. The GPL itself should be licensed under the GPL.

Contrary to what stikonas says, a copyright licence like the GPL is clearly a functional work. Its function is to give users of some software certain freedoms with it, and to let those users know what freedoms they have. No disadvantages, including unimportant arguments like "people might call it 'Linux'" or "people might call it 'open source'" justify taking away everyone's freedom to distribute modified versions of the GPL ("so the whole community benefits" as described in the free software definition.)

The GNU project say that "We recommend making all sorts of educational and reference works free also, using free documentation licenses such as the GNU Free Documentation License (GNU FDL)". I'd say that the GPL is a reference work, so they're being extremely hypocritical, and acting as some form of Lord of the Licences, in a similar way to developers of proprietary software. If I released software, I'd consider not using the GPL for this reason, to avoid giving up my freedom, in a similar way to how I don't use non-free software when I can easily avoid it.

It doesn't matter how you classify copyrightable works (functional, aesthetic, reference). The fact remains that the public's liberty to share or modify them has been suspended.

FSF's concern is freedom. It's focus may be software, but that shouldn't blind it to the application of its principles in other fields of endeavour.

Disallowing modification of the license is simply a crude assertion of the moral right of integrity, i.e. that modified licenses may not be permitted where they are confused as the original work, the work of the same author, or approved/endorsed by the same author, etc.

It may lengthen the licence to achieve this, but it would be more principled to permit derivative licences where they observed the necessary criteria (could not possibly cause confusion).

Comment 3191: verbatim or exact

I see that "verbatim" is used thrice, and "exact" is used once in the license. Can we used either "verbatim" or "exact"

Comment 3192: Why was this not removed with the new draft?

Why was this clause not amended or removed for the final draft? Aside from the patent cut-off date, this was one of the most commented on sections of the previous draft and they unanimously called for removal of this limitation.

I have read the guide, FAQ and rationale of this final draft but not one word is said about why this was left unchanged.

I am still of the opinion that this limitation to thrird parties being in the business of distributing software is a bad thing, considering patent trolls. A software company can offload some IP on a litegation/troll company which in turn could exract a toll from Free Software, just like MS attempted with the MS-Novell deal.

I would appreciate an exaplantion (add it to the FAQ?) why I and the other commenters from the previous draft are wrong and why this section was kept.

Kind regards,

-- Sander Marechal

I am also concerned about this. Even Microsoft has their "Microsoft Licensing" subdivision (mentioned in the Novell-Microsoft deal), which I don't think distributes software. So this clause may not even completely prevent the actual Novell-Microsoft deal, let alone variants.

How does preventing a patent troll from distributing GPLv3 software achieve anything? There's no point in patent trolls being included; this is not a weapon that can hurt them. Instead, forbidding people to license patents from trolls (given that those patents exist) just hurts the free software community. Doesn't it?

gerv:

1. We are not discussing 'preventing a patent troll from distributing GPLv3 software' since the basis of our concern is that the text refers to a company who is NOT in the business of software distribution.

2. The wording appears to disallow deals similar to the Microsft/Novell agreement, since Microsoft IS in the business of distributing software and thus Novell would not have been able to make the agreement that they did.

3. However, under this wording Novell could still make a 'discriminatory' patent deal with a 'troll' (i.e.) a company not in the business of distributing software).

4. This appears to be a major loophole. If it was noted in previous drafts, what is the rationale for it still being included?

Comment 3193: Definition of ROM?

I guess, that this means Read-Only Memory or Read-Only Media. Does this include rewritable ROM too e.g. Flash-ROM or EEPROM (or whatever they are called)? I guess it doesn't, but I might be wrong. This needs to be cleared up.

This issue was originally brought up in the comment 3112, but seems it got ignored.

"Non-rewritable memory" might be better.

Comment 3194: This draft continues to ignore other interactions of executable works

The draft, as do previous drafts, continues to ignore cases in which multiple executable works are somehow combined to form an overall system but not by having works serve as libraries to other works.

For example executable works may exchange data and control signals in some way. Clearly, if a user prepares a file with one work, manually inputs the file to another work, and manually provides the output to the first work, the works can be considered completely independent. However, if the second work is GPLV3-licensed, with what kind of interaction does the first work become part of a "work as a whole" subject to GPLv3 licensing together with the second work?

Comment 3195: The draft continues to inadequately define "the entire work, as a whole"

See my comment under the definition of System Libraries. This draft, as did previous drafts, continues to fail to clearly define what work is covered under the license. It does so for interactions involving system libraries only. It does not take into consideration situations in which multiple executable works coordinate by somehow passing data and control information to form an overall system. An example would be a database application and an analysis program that receives data extracted by the application from the database, processes it, and returns results that the application enters into the database.

This can be one work or it can be two. At what point does it become one work? This draft fails to provide guidance for determining that point. The common interpretations of Version 2 did provide guidance: If the interaction is by means commonly used by separate works to communicate (AFAIK, usually meaning files, pipes, messages, and similar means) then the works can be treated as separate.

This draft somehow needs to make that determination clear.

Comment 3196: Still vague

As noted in a comment on the previous draft, "retains the ability" is unclear - does it mean "is physically able to", "is allowed to", or "is legally permitted to".

I think "can install" would be clearer, especially since the difference between "can" and "may" as "physically able to" versus "is allowed to" is drummed into all native English speakers early on...

Comment 3197: Translations

This license will be translated in other languages, and there are officially bi- or multi-lingual countries [eg Canada]. You should add a clause stating that the English version shall be definitive in the unlikely possibility that a translated version is interpreted differently (legally speaking) from the original.

I already commented about this elsewhere, but I also want to add it here; I think it would be good to also incorporate a clause that states the rights to translate this document. I.e. is it OK to give a translation instead of the english version, when the licenced software is distributed locally - especially when this happens in a country where most people don't speak English? Maybe this could be done with a pointer to the original licence (telling that this one is official, and where to find it). At least this should be made clear so that different interpretations don't evolve...

Comment 3198: why not just make a list

Would it not make sense to simply have an (extensible) list of agreements which are grandfathered in and a process for adding to that list? Anybody who has a fear of the license could simply ask for their agreements to be accepted onto the list before they use the license. As this clause stands it's impossible to know what it means since nobody knows exactly which agreements would be covered.

It will be difficult to maintain and will not protect open source software (It can create a monopoly).

Comment 3199: Where do contributors grant the patent license?

The license makes a *promise* of a patent license by past contributors, but never actually unconditionally *requires* contributors to grant such a license. For example, past contributors may have used alternative (1) in section 11, or may simply deny the "knowingly" in section 11.

So, you're forcing propagators of GPLv3 software to make broad promises on the existence of patent licenses outside their control, and which they have no reason to believe exist. That's an insane liability to accept.

Oddly, this bug was intentionally introduced in a past draft but never explained, despite repeated queries via gpl3.fsf.org. What's going on?

The alternatives (1, 2, 3) are for conveyors, not contributors. Conveyors don't have to grant a patent license, but section 11 tries to stop them from relying on a discriminatory license. I have been wondering (even during the previous draftss) about your original point, that the license says contributors must grant a patent license, but doesn't actually make them. IANAL, but I've now concluded that because contributors agree to the whole license ("You must license the entire work, as a whole, under this License") they /probably/ must license their patents as this clause describes. However, it could be made much more explicit by adding the requirement to Section 5 (between c and d, "You must grant to the recipients and all parties a non-exclusive, worldwide, royalty-free patent license under the your essential patent claims, to make, use, sell, offer for sale, import and otherwise run, modify and propagate the contents of the contributor version."). Of course, contributor version and contributor should then be defined earlier (ideally in Definitions).

Yes, you're right. When a contributor distribute the work she must do so under the full license, which includes her own granting of a patent license. Good.

Still, as a distributor I'm worried about making a representation about licenses from all *past* contributors. If there *was* a copyright violator somewhere upstream (and usually I have no way of telling for sure) , this seems to place me in considerable legal risk.

Comment 3200: Disallow service conditioned on voluntary locking

Comment 2925 is still pertinent. Why is it ignored in this draft?

Repeating that comment:

One way to circumvent the anti-tivoization is as follows: the device's vendor allows installation of modified works by default, but has a "vendor lock" flag that can be enabled by the user and prevents such modifications. Service is then conditioned upon the user "voluntarily" setting the flag. See my comment 2923 for a full description.

This is a potentially devastating loophole: the user gets the choice of a free but useless device, or a serviced but forever-unmodifiable device.

To close this loophole, I suggest adding a sentence to this paragraph, right after the first sentence:

"However, support service, warranty, and updates must not be conditioned upon acceptance of a constraint or obligation to avoid future modifications or installation of modified works."

While this comment might make sense with regard to the following sentence of the license (about network access), it makes no sense with regard to this one. Anti-TiVoization is not about "support service, warranty, and updates", all of which it makes complete sense to condition on use of a particular version of the software. If you modify the software in your TiVo, you might or might not have an expectation of it going on working with the network, but you have NO expectation that TiVo will provide support service (i.e., customer support), warranty, or updates.

Sure, I don't expect support after I modify the software. But this issue is about conditioning support in the *present* on an obligation/constraint to never modify the future in the *future*. This doesn't seem justified, and creates a big loophole allowing vendors to force users to "willingly" give up their rights.

Hmm. I guess I would have said that it was OK for someone offering to provide "support service, warranty, and updates" to predicate that offer on anything they wanted - that it's just outside the scope of the license, even if it happens to be an offer from the one who provided the software to begin with. This assumes that the device/software is operational without the offered support and the user has a real choice as to whether to accept the offer or not.

I think this clause was intended to address onlly the narrow question of whether the vendor could terminate support, warranty, etc. if the software was found to be modified.

Comment 3201: Independently actionable procedure?

Comment 2488 remains unaddressed, despite being acknowledged by Brett Smith.

Briefly, what's to prevent these procedures from including interaction with a vendor, or relying on a $10,000 device-locked SDK?

Comment 3202: Style: Program or Work?

Since the GPL is designed to be applicable to any work of authorship (not only computer programs), I once again suggest using a more neutral term than "the Program". Something like "the Work" would avoid misleading many many people into thinking that the GPL can only be applied to computer programs.

I agree, somewhat, but I do make note that they allow for other things to be licensed. Direct quote: "The Program" refers to any copyrightable work licensed under this License.

But I agree, "The Work" could be a bit helpful, compared to "The Program"

I agree that using "Work" everywhere is clearer and more effective. BTW, I think there should be no article, especially not in quotes! (Or one may think that the definition only applies to "The Work", not to "Work" or "work")

It's also worth noting that despite the term "Program" being defined in the license itself, there can still be cases where it isn't clear what the license applies to; this was asked recently on one of the Doxygen mailing lists: a user noticed that, while the source code for Doxygen itself had the relevant license notices, the code for Doxywizard, instead, which is usually distributed *together* with Doxygen, didn't have any. So, he asked, "is Doxywizard distributed under the GNU GPL?". I think the question is legitimate: the technical meaning of "program" leads to identifying separate programs in the distribution (Doxygen, Doxytag, Doxywizard...); whether the same term, when used in the license notices, refers to one of these programs or the the whole "work" being distributed isn't clear at all AFAICS.

Comment 3203: Good: clear definitions

The definitions of "modify", "modified version", work "based on" another work, and "covered work" are unchanged with respect to GPLv3draft3, and they are very clear. It's good that the definition of "modified version" exploits applicable copyright law without trying to rewrite it. This ensures that the license does not place restrictions on activities that do not require permission under applicable copyright law.

"This ensures that the license does not place restrictions on activities that do not require permission under applicable copyright law."

IMO the main problem is, that when distributing software in the internet, every copyright law in every country could be applicable, depending on who downloads the software. So if someone develops (closed source) software in country A, the court in country A could say, that the software is not "based on" covered work. But if this software is available on a website and downloadable from country B, courts in Country B might say, that it's based on covered work an the source code has to be disclosed. This problem does not exist with GPLv2, as US-legal-terms are used and the courts in every country interpret them the US-way and not in their own way.

ma: I do not think that the use of US-centric terms in the GNU GPL v2 can insure that every country applies USC definitions. There's no choice of law in the GNU GPL v2, there's just use of US-centric terminology in a license which is supposed to be applicable in every country.

Denationalizing the terms in the GNU GPL v3 is an improvement, IMO.

Comment 3204: Good: clear definitions

The definitions of "propagate" and "convey" seem fairly clear. Again, their linking to copyright law ensures that the license does not place restrictions on activities that do not require permission under applicable copyright law.

Comment 3205: Good: useful definition

This definition is really useful to simplify the language of clause 5d (even though I would like to see clause 5d dropped entirely, more on this in a later comment...).

Comment 3206: Good: clear and appropriate

This section seems OK to me.

Comment 3207: Good: please keep these definitions

I like the definitions of "source code" and "Object code" as they are. Good to see that they were kept unaltered.

Comment 3208: Good: mentioning of fair use is an improvement

Good to see that fair use and similar rights are explicitly mentioned. This is an improvement over GPLv2.

Comment 3209: Problematic: possibly untrue

This clause is unchanged with respect to GPLv3draft3, and thus still problematic. It could be seen as an untrue statement in some cases. How can the licensor say that the covered work won't be judged as "part of an effective technological measure" under a given law? That is for the courts to decide. In some scenarios, GnuPG may actually be considered "part of an effective technological measure" and could be deemed so by a judge...

Let's say that someone develops a new DRM system and releases it under GPLv3-draft4. A movie studio starts using it when distributing movies over the Internet. Someone takes the GPL'd DRM code and modifies it so, that it removes the DRM from the downloaded movies. Now the movie studio cannot (sanely) sue solely because the DRM was cracked, because the DRM system the movie studio had used was under GPLv3-draft4.

Similarly, if copyright holder had encrypted a work with GPLv3'd GnuPG, the copyright holder couldn't reasonably claim that "circumventing" such protection was against DMCA.

Maybe I'm missing something, but this is how I understand this section.

If the copyright holder is also distributing the implementation of the DRM or encryption scheme, this claim might work. If the implementation is part of a player manufactured independently, or if the copyright holder simply USED GnuPG to encrypt and the decryption in players was provided by the manufacturer, then the copyright holder would not be a party to the license and could not be said to have implicitly allowed circumvention.

Comment 3210: Bad: possibly overreaching

This clause is unaltered with respect to GPLv3draft3, and thus still troublesome, as it seems to be overreaching. For instance, it could be interpreted as covering legal powers to forbid "computer crimes" such as unauthorized intrusion into computer systems.

E.g.: suppose that the covered work is a vulnerability scanner, or password cracker, or anyway a tool that could be used (among other things) to break into other people's computers. Using that tool in this manner is exercising a right "under this License" and is a circumvention of appropriate technical measures set to protect a computer system or network from unauthorized access. Gaining unauthorized access to a protected computer system or network is forbidden by law in several jurisdictions; do I waive such a legal protection, when I convey the covered work?

I suggest dropping the waiver entirely, thus leaving the following disclaimer only.

Amend it: “you waive any legal power to forbid circumvention of technical measures... by a party on a device legitimately in their own possession”

Comment 3211: Good: substantially unchanged

This section is unchanged with respect to GPLv3draft3. It's not that dissimilar to the corresponding section in GPLv2 (apart from the references to section 7) and looks OK to me.

Comment 3212: Good: improved wording

This subsection is clarified as to which date(s) must be given. This is an improvement, since many people were uncertain on how the corresponding language in GPLv2 should be interpreted.

Comment 3213: Good: disjunctive licensing is explicitly acknowledged

I think that explicitly acknowledging the possibility of dual licensing (GPLv3 or OTHERLICENSE at the recipient option) is good, as it clarifies that the GPL does not try to take away permissions that are separately granted. This is an improvement over GPLv2: good to see that it was kept in GPLv3draft4 too.

Section 5 seems OK, so far (apart from the reference to Section 7, more on this in a later comment...).

Comment 3214: Bad: too restrictive

Clause 5d is now simpler and clearer than in the previous drafts: as a consequences, its issues are more apparent! ;-)

This clause is worse than the corresponding clause 2c in GPLv2... :-(

It's an inconvenience and border-line with respect to freeness. Actually this clause restricts how I can modify what an interactive program does when run. It mandates a feature that I *must* implement in *any* interactive interface of my modified work. It's very close to place an unacceptable restriction on modification. What is more awkward is that it seems that when a non-interactive work is modified so that it becomes an interactive work, the modifier is *compelled* to implement these features in *any* newly created interactive interface...

I would like to see clause 5d dropped entirely.

Or, at least, it could be modified so that it only applies to cases where the original Program is also interactive. Something like:

| d) If the Program has interactive user interfaces which display | Appropriate Legal Notices, this feature must be preserved in each | interactive interface that is also present in the work.

Yes. It is one thing to require preservation of the Appropriate legal notices feature in existing interactive user interfaces.

It is entirely different to compel users to include such a feature in any newly created interactive interfaces. This is far worse than the equivalent clause in GPL v2. People WILL ignore this requirement, and assume it merely mandates preservation of the feature in existing interfaces. This is a *critical* problem with the license IMHO.

Placing notices on interfaces is an important way of informing users of their rights. Most people don't read the source code.

Comment 3215: Good: basically unchanged so far

Section 6 seems so far to be basically similar to section 3 of GPLv2, just more explicit and clear in some respects and a little more permissive in other ones. Good.

Comment 3216: Good: improved definition which is no longer U.S.-centric

This part is improved with respect to GPLv3draft3, as it no longer refers to U.S.-specific acts. Good.

It's better because it's simpler, IMHO.

Comment 3217: Good: clarified requirement to accompany source with installation keys

This requirement to provide Installation Information is basically unchanged with respect to GPLv3draft3: it's a reworded and clarified version of the requirement to include installation keys in the Corresponding Source (which was present in GPLv3draft2). The current language seems clear. Good.

Comment 3218: Kills copyleft: this is not in the spirit of GPLv2

I strongly *dislike* the entire concept of allowing a limited set of additional requirements to be added.

That is *against* the spirit of the GPLv2 (where the FSF promised that new versions would "be similar in spirit to the present version", see GPLv2, section 9.) and greatly weakens the copyleft.

Yes, it enhances license compatibility (as the rationale of GPLv3draft1 explains), but the GPL never was a champion of compatibility. If the goal were maximizing license compatibility, the best choice would be adopting the 2-clause BSD license (http://www.gnu.org/licenses/info/BSD_2Clause.html) as the new GPL... It would be compatible with pretty everything and clearly free, even though non-copyleft. I'm *not* against non-copyleft free licenses, but they definitely aren't "similar in spirit to" GPLv2...

I would be happy to see all these "permissions to add restrictions" entirely dropped from Section 7.

Comment 3219: Kills copyleft: are these the cousins of GFDL's Invariant Sections?

What exactly is a "reasonable legal notice"? What exactly is an "author attribution"? It seems that these terms are not defined anywhere in the license. I'm concerned that they could be interpreted in a broad sense and allow people to take a GPLv3'd work and add some sort of invariant long text that nobody will ever be able to remove or modify... This option could make a work include unmodifiable & unremovable parts and thus fail to fully grant the freedom to modify. I would rather avoid introducing such options in the GPLv3!

While I'm sure that it's not FSF's intention, I think this could be also read to allow BSD advertising clause. People like to use OpenSSL in GPL'd projects, so I wouldn't be surprised if several programmers would think that GPLv3-draft4 is compatible with the license of OpenSSL (they would think that "wasn't GPLv3 about improving license compatibility?").

This section is too unclear indeed.

I agree with the earlier posters. Allowing any kind of "author attributions" is far too broad. The Open Source Initiative has been considering several licenses that demanded onerous forms of attribution, including hyperlinks and logos that must be maintained exactly (even specifying an exact size). For example, consider the SocialText Public License (http://www.socialtext.net/open/index.cgi/spl100.pdf?action=attachments_download;page_name=socialtext_open_source_wiki;id=20060725230840-0) It says, "all copies of the Covered Code in Executable and Source Code form distributed must, as a form of attribution of the original author, include on each user interface screen (i) the "Powered by Socialtext Open Source Edition" logo and (ii) the copyright notice in the same form as the latest version of the Covered Code distributed by Socialtext, Inc. at the time of distribution of such copy. In addition, the "Powered by Socialtext" logo must be visible to all users and be located at the very bottom center of each user interface screen. Notwithstanding the above, the dimensions of the "Powered By Socialtext" logo must be at least 100x25 pixels. When users click on the "Powered by Socialtext Open Source Edition" logo it must direct them back to http://www.socialtext.net/stoss. In addition, the copyright notice must remain visible to all users at all times at the bottom of the user interface screen. When users click on the copyright notice, it must direct them back to http://www.socialtext.net/stoss." This kind of complex graphical attribution drastically limits modification. Among other things, it constrains GUIs, completely forbids text-only interfaces, and requires internet access; it is also essentially impossible to satisfy more than one of these clauses at a time. I think the allowed forms of attribution should be strictly limited to avoid allowing this type of clause, even if it hurts compatibility somewhat.

What about adding some examples of legal notices that could be reasonable or not? This paragraph could contain someting like: "(reasonable legal notices could include short aknowledgments of other works on which this work is based or grants that sponsored the development of this work, but not include propaganda for or against any cause or company)". For the latter, I am thinking about the cdrecord fiasco and the claim that some propaganda against Linux could not be removed from the code because it was part of the copyright statement protected by the GPL.

'Copyleft' requires some restrictions otherwise the GPL would not be an effective at protecting all users' rights under the GPL.

1. COMPARISON BETWEEN GPL V3 DRAFT AND GFDL.

Neither section 7b nor any other part of section 7 is similar to the provision for Cover Texts and Invariant Sections in the GFDL. The terms of the GPL V3 draft including section 7 are much more similar to the terms of the GSFDL draft which disallow GFDL Cover Texts and Invariant Sections. Although, one should not expect a perfect analogy between a software license and a documentation license.

1.1. ORIGIN OF GFDL.

GFDL had included provision for Cover Texts and Invariant Sections as an inducement to publishers to create high quality free documentation for free software. There is still a shortage of free documentation of the highest quality for free software, although, some forms of free documentation have been introduced by some publishers specialising in computer related titles.

1.2. GSFDL.

I expect to use the GSFDL for documenting my own work once the GSFDL draft period is over and the GSFDL is released.

2. REASONABLE LEGAL NOTICES.

It would be difficult to define reasonable legal notices in an especially careful manner which would also be a helpful manner. Reasonable legal notices has an adequately well defined meaning under law and in ordinary language. Reasonable notices are not unreasonable notices. Legal notices are not notices which are not of a legal nature. The license cannot list every possible legal notice that may needed to satisfy what may be required to convey every type of software for every local jurisdiction in the world.

A legal notice may have a statement of purpose, however, no reasonable legal notice or any type of copyright notice can properly include a diatribe for or against a particular operating system or operating system kernel.

You should not have difficulty identifying a notice which is not a reasonable legal notice and removing it if you choose.

3. AUTHOR ATTRIBUTIONS.

Author attributions have a value to users comparable in importance to copyright notices in a world of copyright assignment and orphan works. I will not repeat the full arguments from my comment on the previous draft, here in this comment. Those arguments may be read at http://gplv3.fsf.org/comments/rt/readsay.html?id=3140 .

Author attributions are certainly textual information which identifies authors and the portion of the code for which they are responsible. This would reasonably include the name and contact information for an author. John Smith may believe that his name alone is a sufficient identifier for his name and if he requires nothing else that is fine. However, to usefully identify himself as distinct from all the other John Smiths of the world, he, may choose to add identifying information such as his postal address, his affiliation, his email address, and the URL to his website. If a user is lucky, some years after the copyright holder goes out of business, John Smith has moved, and changed the affiliation which had provided his email address a user may still be able to identify John Smith via his URL or some other identifier which may not have changed. If the user is able to identify John Smith, the user may then have an opportunity to contact the author and obtain the author's own answer to how to best understand the intent behind some poorly commented part of the code and how to best modify some code to introduce some desired change.

A logo or a photograph of the author crosses the line as useful for identifying the author at a distance. The example of the SocialText Public License 1.0 cited by a previous child comment would not be compatible with the GPL V3 drafts on several grounds even if it incorporates some text which has appeared in GPL V3 drafts. Specifying a required size in pixels is silly because the resolution of the user's display is unknown and would therefore not reliably achieve the intent. The language used in Appropriate Legal Notices definition, "a convenient and prominently visible feature" and "a prominent item in the list", addresses the same intent in a simple reliable manner which also works for command line interaction.

Information to identify an author is perfectly appropriate. If a user chooses to contact an author by post, email, URL, or whatever; the author may include a logo or a diatribe for and against something as part of his response to the user. That would be the author's free speech right when someone requests communication with him. However, no requirement to include identifying addresses when they are part of a required attribution would compel you to also include messages from the author. If an author has included such messages accompanying an attribution, you are well within your rights to remove such messages. If you choose to leave such messages, you would only be doing so as a mere courtesy to the author. You may exercise your choice on the question when you convey modified versions.

You should not have difficulty identifying information which is not properly part of an author attribution and removing it if you choose. 3. BSD ADVERTISING CLAUSE.

No terms complying with this section could resemble the advertising clause of the original BSD license. That clause required including a statement of acknowledgement in advertisements mentioning the program. The GPL V3 drafts have never had any terms which would compel the inclusion of any statement in a context independent from the program itself. No such terms would ever be considered reasonable legal notices.

All of the additional restrictions allowed in section 7 for limiting the publicity use of licensors names; or the use of trade names, trade marks, or service marks are given in a negative form. No provision for such terms should be presumed to compel statements. Compliant terms would merely remind users that appropriating someone else's name or mark is not allowed. People should know better in any case, but I have no problem if someone wants to remind users about what they are already not allowed to do with someone else's name or mark. 4. LIMITATIONS ON FORM OF PRESENTATION.

Concerns about improper specification of the form of legal notices and the form of attributions should be satisfied by restricting the notices and attributions to "in that material or in the Appropriate Legal Notices displayed by works containing it".

This language should be understood in a manner which would not allow arbitrary specification of the size of the legitimate elements in pixels or other inappropriate specifications. Some local laws do require some legal notices to have a minimum font size, however, I am uncertain to what degree those local laws might have a practical effect on computer displays when font size is not really controllable by many programs. I am not concerned about legal notices myself so I would have to plead ignorance on the font size issue for some local laws.

I have suggested a correction to the inconsistency with the definition of Appropriate Legal Notices in section 0 and also for improving how Appropriate Legal Notices are applied in section 7. See http://gplv3.fsf.org/comments/rt/readsay.html?id=3378 .

If the font size issue which I raised for legal notices is moot for software displays, then I would suggest considering a clarification of the form of presentation. An additional sentence in section 7 might provide something like the following clarification.

"No non-permissive terms added in accord with this section may specify the form of presentation for non-permissive terms in a more restrictive manner than requiring their inclusion in the relevant source files and in the display of Appropriate Legal Notices if the work displays Appropriate Legal Notices."

The alternate form of clarification avoiding excessive visual centricism would be as follows.

"No non-permissive terms added in accord with this section may specify the form of presentation for non-permissive terms in a more restrictive manner than requiring their inclusion in the relevant source files and in the Presentation of Appropriate Legal Notices if the work Presents Appropriate Legal Notices."

Sorry, my previous post had not been the last draft of my comment which follows.

'Copyleft' requires some restrictions otherwise the GPL would not be effective at protecting all users' rights under the GPL.

1. COMPARISON BETWEEN GPL V3 DRAFT AND GFDL.

Neither section 7b nor any other part of section 7 is similar to the provision for Cover Texts and Invariant Sections in the GFDL. The terms of the GPL V3 draft including section 7 are much more similar to the terms of the GSFDL draft which disallow GFDL Cover Texts and Invariant Sections. Although, one should not expect a perfect analogy between a software license and a documentation license.

1.1. ORIGIN OF GFDL.

GFDL had included provision for Cover Texts and Invariant Sections as an inducement to publishers to create high quality free documentation for free software. There is still a shortage of free documentation of the highest quality for free software, although, some forms of free documentation have been introduced by some publishers specialising in computer related titles.

1.2. GSFDL.

I expect to use the GSFDL for documenting my own work once the GSFDL draft period is over and the GSFDL is released.

2. REASONABLE LEGAL NOTICES.

It would be difficult to define reasonable legal notices in an especially careful manner which would also be a helpful manner. Reasonable legal notices has an adequately well defined meaning under law and in ordinary language. Reasonable notices are not unreasonable notices. Legal notices are not notices which are not of a legal nature. The license cannot list every possible legal notice that may needed to satisfy what may be required to convey every type of software for every local jurisdiction in the world.

A legal notice may have a statement of purpose, however, no reasonable legal notice or any type of copyright notice can properly include a diatribe for or against a particular operating system or operating system kernel.

You should not have difficulty identifying a notice which is not a reasonable legal notice and removing it if you choose.

3. AUTHOR ATTRIBUTIONS.

Author attributions have a value to users comparable in importance to copyright notices in a world of copyright assignment and orphan works. I will not repeat the full arguments from my comment on the previous draft, here in this comment. Those arguments may be read at http://gplv3.fsf.org/comments/rt/readsay.html?id=3140 .

Author attributions are certainly textual information which identifies authors and the portion of the code for which they are responsible. This would reasonably include the name and contact information for an author. John Smith may believe that his name alone is a sufficient identifier for his name and if he requires nothing else that is fine. However, to usefully identify himself as distinct from all the other John Smiths of the world, he, may choose to add identifying information such as his postal address, his affiliation, his email address, and the URL to his website. If a user is lucky, some years after the copyright holder goes out of business, John Smith has moved, and changed the affiliation which had provided his email address a user may still be able to identify John Smith via his URL or some other identifier which may not have changed. If the user is able to identify John Smith, the user may then have an opportunity to contact the author and obtain the author's own answer to how to best understand the intent behind some poorly commented part of the code and how to best modify some code to introduce some desired change.

A logo or a photograph of the author crosses the line as useful for identifying the author at a distance. The example of the SocialText Public License 1.0 cited by a previous child comment would not be compatible with the GPL V3 drafts on several grounds even if it incorporates some text which has appeared in GPL V3 drafts. Specifying a required size in pixels is silly because the resolution of the user's display is unknown and would therefore not reliably achieve the intent. The language used in Appropriate Legal Notices definition, "a convenient and prominently visible feature" and "a prominent item in the list", addresses the same intent in a simple reliable manner which also works for command line interaction.

Information to identify an author is perfectly appropriate. If a user chooses to contact an author by post, email, URL, or whatever; the author may include a logo or a diatribe for and against something as part of his response to the user. That would be the author's free speech right when someone requests communication with him. However, no requirement to include identifying addresses when they are part of a required attribution would compel you to also include messages from the author. If an author has included such messages accompanying an attribution, you are well within your rights to remove such messages. If you choose to leave such messages, you would only be doing so as a mere courtesy to the author. You may exercise your choice on the question when you convey modified versions.

You should not have difficulty identifying information which is not properly part of an author attribution and removing it if you choose.

4. BSD ADVERTISING CLAUSE.

No terms complying with this section could resemble the advertising clause of the original BSD license. That clause required including a statement of acknowledgement in advertisements mentioning the program. The GPL V3 drafts have never had any terms which would compel the inclusion of any statement in a context independent from the program itself. No such terms would ever be considered reasonable legal notices.

All of the additional restrictions allowed in section 7 for limiting the publicity use of licensors names; or the use of trade names, trade marks, or service marks are given in a negative form. No provision for such terms should be presumed to compel statements. Compliant terms would merely remind users that appropriating someone else's name or mark is not allowed. People should know better in any case, but I have no problem if someone wants to remind users about what they are already not allowed to do with someone else's name or mark. 5. LIMITATIONS ON FORM OF PRESENTATION.

Concerns about improper specification of the form of legal notices and the form of attributions should be satisfied by restricting the notices and attributions to "in that material or in the Appropriate Legal Notices displayed by works containing it".

This language should be understood in a manner which would not allow arbitrary specification of the size of the legitimate elements in pixels or other inappropriate specifications. Some local laws do require some legal notices to have a minimum font size, however, I am uncertain to what degree those local laws might have a practical effect on computer displays when font size is not really controllable by many programs. I am not concerned about legal notices myself so I would have to plead ignorance on the font size issue for some local laws.

I have suggested a correction to the inconsistency with the definition of Appropriate Legal Notices in section 0 and also for improving how Appropriate Legal Notices are applied in section 7. See http://gplv3.fsf.org/comments/rt/readsay.html?id=3378 .

If the font size issue which I raised for legal notices is moot for software displays, then I would suggest considering a clarification of the form of presentation. An additional sentence in section 7 might provide something like the following clarification.

"No non-permissive terms added in accord with this section may specify the form of presentation for non-permissive terms in a more restrictive manner than requiring their inclusion in the relevant source files and in the display of Appropriate Legal Notices if the work displays Appropriate Legal Notices."

The alternate form of clarification avoiding excessive visual centricism would be as follows.

"No non-permissive terms added in accord with this section may specify the form of presentation for non-permissive terms in a more restrictive manner than requiring their inclusion in the relevant source files and in the Presentation of Appropriate Legal Notices if the work Presents Appropriate Legal Notices."

/ Reasonable legal notices has an adequately well defined meaning under law and in ordinary language. Reasonable notices are not unreasonable notices. /

I'm not sure about this. If 8 lines of text is reasonable, then surely 12 lines is reasonable, and if 12 lines is reasonable, then a page is surely reasonable... Also, something that might be reasonable in the original work, might not be reasonable in a derived work adapted for mobile devices. If this is to be included, I think there should be some per-contributor limit.

Comment 3220: Good: further restrictions are void

I'm glad to see that this is explicitly stated: every attempt to license a work under the terms of GPLv3 with further restrictions is equivalent to licensing under the plain GPLv3. This is good, since there are unfortunately many people that license works in inconsistent manners (such as GPLv2 + additional restrictions); creating a rule that resolves this kind of inconsistencies for the better is a good thing to do.

Comment 3221: Improvable: fails to mention fair use

Section 8 seems to be more or less similar to section 4 of GPLv2, except that it's more forgiving to first-time violators. This is good. However, as in GPLv2, it fails to mention fair use and similar rights and possible separate licenses (as in disjunctive dual-licensing...). It should be improved to acknowledge these possibilities.

Comment 3222: Improvable: again fails to mention fair use

Another section that fails to mention fair use and similar rights and possible separate licenses. Section 9 should be improved by acknowledging these possibilities.

Fair use is acknowleged earlier in the license. Why make it longer by repeating oneself?

gerv: just for clarity's sake...

In the case of dual-licensed software, this section makes a blatantly false claim. That's a bit poor for a legal document. It looks like it's been copied from an EULA (where it would be true), without being edited.

It should say something like: "If you do not otherwise have permission to propagate or modify a covered work, then such an action would infringe copyright if you did not accept this License. Therefore, by modifying or propagating a covered work without otherwise having permission to do so, you indicate your acceptance of this License."

Comment 3223: Good: patent litigation prohibition doesn't seem to be overreaching

This clause forbids patent litigation in a manner that seems to be narrow enough to be free: it only forbids patent suits related to the Program and only concerns people who need the License to convey the covered work. Looks OK.

Comment 3224: Bad: possibly not enough as a protection against patents

I don't quite understand how (1) can be seen as a specific form of shielding downstream recipients. If I am a downstream recipient who does not have a patent license, what protection (against patent infringement lawsuits) would I get from the existence of a network server which makes source available to the public?

I'm puzzled.

if the corresponding source is "so available", it means it is gpl'd, granting you a "non-exclusive, royalty free, worldwide patent license". it stays unclear, though, which steps are necessary for you to become a user of that program (your program may conflict with a patent w/o being a derived of the "corresponding source")

webograp: I'm not following you, I'm afraid... :-(

The "non-exclusive, worldwide, royalty-free patent license" is granted by contributors. But in the case under consideration, a distributor (who is not necessarily a contributor) is conveying the work, knowingly relying on a patent license granted to him/her by a third-party patent holder (who is not necessarily a contributor). As a consequence, the mere fact that the distributor causes the source to be available from a public server under the terms of the GPLv3, does not seem to enhance the situation for me: if I receive the work and I do not have a patent license, I do not feel any more protected from patent infringement lawsuits...

Still puzzled, unfortunately.

I guess the idea is that a patent is breached when a covered device is assembled, and conveying is being taken as the equivalent of assembly for intangible things? That's my guess.

Comment 3225: Good: this is a no-discrimination clause

This clause seems to be intended to prevent the negative consequences of Microsoft-Novell-like deals. It seems to be well drafted, since it's based on a no discrimination principle: you cannot grant a patent license to some recipients only; if you grant a patent license to someone, you automatically extend it to every other recipient. Good.

related to the work and derived works, not a general patent license, hmm?

Comment 3226: Good: this is a no-further-restriction-through-patents clause

This clause seems to be based on the general principle that no further restrictions can be added beyond the terms of this License. In the present case, it is stated that one cannot add further restrictions through restrictive patent licenses. Looks OK.

Comment 3227: Bad: the date is arbitrary

Why this date? What's special about it? Is it just the date of publication of the GPLv3draft3? Was it chosen in order to avoid placing retroactive restrictions, by chance? If this is the reason, I don't think this date limit is needed: since the GNU GPL v3 is not yet released, no work has been licensed under its terms yet, and hence no provision can be retroactive...

The previous draft rationale explained that this date limit is intended to avoid forcing companies to face problems due to past agreements that cannot be changed anymore. I'm not convinced that this would be unfair: agreements like those are bad things and hence making them incompatible with conveying GPLv3'd works does not seem to be unfair.

Patent agreements may well be bad, but the people who entered into them might not have known that. And even if "ignorance of the law is no excuse", there wasn't any law against it until now, so it /would/ be unfair to punish them.

Comment 3228: Good: basically unchanged

This section is identical to section 12 of GPLv3draft3, and thus seems substantially the same as section 7 of GPLv2, apart from minor details. I think it's OK.

Comment 3229: Kills copyleft: compatibility with a yet unknown license

This section introduces a form of compatibility with a license that is yet unreleased and thus possibly non-free: the GNU Affero General Public License, version 3. The AfferoGPL v1 is, in my opinion, a non-free license, due to its clause 2d. I won't restate all the reasons for my conclusions (more details in http://gplv3.fsf.org/comments/rt/summarydecision.html?filename=gplv3-draft-2&id=1663). As a consequence, I have few hopes that the forthcoming version 3 of the GNU AfferoGPL will be a free license.

Being compatible with an unknown (and thus possibly non-free) license destroys the copyleft mechanism of the GPLv3. The previous draft rationale stated that most other free software licenses also feature such form of compatibility with the AfferoGPL: this is true for non-copyleft licenses, though! When I choose a long and complicated copyleft license such as the GNU GPL, I want a copyleft mechanism that actually works; or otherwise I can choose short and simple non-copyleft licenses, such as the Expat or the 2-clause BSD ones...

I strongly recommend dropping section 13 entirely.

Being compatible with an unknown (and thus possibly non-free) license destroys the copyleft mechanism of the GPLv3.

That simply isn't true. This is a linking exception, not a code-mixing exception. GPLv3 code still has to be dealt with under GPLv3 terms. This doesn't destroy its copyleft.

The documents released with this draft show that the AGPLv3 is GPLv3 with this clause being instead a "hand over source over the network" clause - and written using the new text from previous drafts, which is not (to my mind) a restriction on modification, because it doesn't define _how_ you have to meet the obligation of offering a download. But whether that's true or not, this compatibility does not "destroy the copyleft".

It is essential to retain the Affero section 1) to provide free software proponents a straightforward way, in a standard license, to spread copyleft for their software which runs as a service, should they choose-- especially as such software comes to dominate, and 2) to allow GPLv3 code to be compatible with such code, or GPLv3's impact and use will lessen.

I can understand concerns about referencing a license that has not been published, but trusting the FSF's motives on this is easy and logical, and yesterday the AGPL draft was posted, so concerns can now be directly assuaged.

I know frx and some others oppose any SaaS-motivated clause, and they've won: it's not in the GPLv3, and allowing compat with AGPLv3 doesn't undermine GPLv3 or copyleft at all. frx continues to spread total disinformation on this topic. Why not declare victory and be happy? You don't have to use AGPL and it doesn't affect GPLv3 code.

gerv: I think we can agree that the LGPL has a weaker copyleft mechanism than the GPL, since it allows linking with non-free code, under certain conditions, even though it does forbid mixing with non-free code.

If you agree with me on this statement, you should be able to see why I think that allowing linking GPLv3'd code with AGPL(v3_or_later)'d code greatly weakens the copyleft mechanism of GPLv3. It actually turns it into an LGPL-like weaker copyleft, even though with respect to only some other licenses (the AGPL v3 and *any later* versions).

lepreau: Spreading copyleft to network software is a good thing to do, *if and only if* it's done without sacrificing the freeness of the license. I have yet to see a clause that closes the so-called "ASP-loophole" without introducing non-free restrictions: I'm not saying that it's impossible to draft one, but, as a matter of fact, I've never seen one so far.

The AGPLv3-draft1 has been published, but please note that this clause introduces a linking permission with works under the AGPL version 3 *or any later version*: hence, my concerns about linking compatibility with yet unknown licenses are still valid.

Finally, I'm not spreading disinformation: as I said in the previous comment (when replying to gerv), the copyleft mechanism is weakened be a linking permission: when I want to allow linking my code with works under arbitrary (and yet unknown) license terms, I use the LGPL (or a non-copyleft license), not the GPL...

Comment 3230: Good: it should be kept as it is

This section is almost unchanged from previous drafts, and similar to section 9 of GPLv2 (apart from minor rephrasing and the addition of the final paragraphs). It's good that licensors have the option of specifying a certain version of the GNU GPL "or any later version", but are not forced to do so. In other words, the upgradeability of the license is a permission that is granted outside the license text and this is a good feature to keep.

Comment 3231: Good: this section looks OK

This section is identical to the first paragraph of section 15 of GPLv3draft3. It looks OK to me.

Comment 3232: Good: this section seems fine

This section is identical to the second paragraph of section 15 of GPLv3draft3. It seems fine to me.

Comment 3233: Good: this section is also a good thing to have

This section is identical to the third paragraph of section 15 of GPLv3draft3. I think it's a very useful thing to have in the license.

Comment 3234: Why 60 days for notification?

So what exactly happens if the copyright holder does not notify the offender before 60 days have elapsed?

This sounds like a "get-out-of-jail-free" card to me...correct me if I'm wrong.

Well, it's 60 days since "the most recent violation", so it's only get-out-of-jail-free if the offender has already stopped violating the license, which pretty much makes the issue moot anyway.

I cannot understand this either. This sounds like that you can completely legally do whatever you like with code under GPLv3-draft4 (and earlier drafts) if copyright holder doesn't "put you on notice [...] provided 60 days have not elapsed since the most recent violation".

In practice, it usually takes several months before a violation is noticed. Does this section automatically forgive most GPLv3 violations simply because people cannot put violator in notice quickly enough?

Example: A company decides to create a router-firewall and use GPLv3'd software on it. They decide to ignore GPLv3 terms and simply install customized binaries to the device without giving anyone source code. They produce one big batch of devices and sell them all to resellers in a short period of time.

(1) Person buys such device from a reseller. Three months later he notices that the device has GPL'd code but no source code was given to him. Does he have any chance to get the Corresponding Source, since "so much" time has already passed?

(2) If one of the resellers notice the violation three months after they have got the devices from the manufacturer, do they have any chance to get the Corresponding Source? If not, is the only legal option to trash the devices, if they don't know how to replace the software in the devices?

(3) The copyright holders hear about this three months later. Can they revoke the license of the manufacturer so that it cannot legally repeat this a year later?

Sorry about long post. It just sounds scary if violations get very easily automatically forgiven, because it makes the license useless. I hope that I have misunderstood this section.

In larzhu's example, if the manufacturer is continuing to sell the device with no means to get the source code, then each such sale is a violation of the license and resets the 60-day clock.

Note that the only recourse defined by the license is termination of the rights under the license, including the right to distribute. If the manufacturer has already stopped distributing the program, then terminating the license wouldn't have any further effect, regardless of the 60-day limit - the only result of that termination is to force them to stop distributing.

One imponderable here is the question of which section 6 clause would apply. The manufacturer might claim that it was distributing under 6a (accompanied by source code) and failed to provide the required source (a violation at the time of distribution, already more than 60 days in the past). On the other hand, the reseller could claim that since the source wasn't provided, one of the other clauses must apply, in which case failure to provide the source on request would presumably be a new violation. However, again, the only thing that violation could trigger would be termination of the right to distribute, which wouldn't help the reseller, but would resolve your (3) question by not allowing the manufacturer to do it again.

Of course, the reseller might have recourse under contract or commerce law, especially if the devices for some reason became unsellable., depending on the terms under which they bought the devices from the manufacturer.

Whether the devices would become unsellable might become an interesting legal case if the copyright holders tried to stop such sales. First Sale might or might not allow the reseller to sell them (which doesn't normally require copyright holder permission, in many jurisdictions), but a copyright lawsuit might get them declared contraband. The final paragraph of the Termination section, which shields downstream recipients of copies distributed in violation of the license, might argue against such an outcome (assuming the resellers weren't colluding with the manufacturer).

Thanks sepreece, your comment clarified this notably to me.

It's still unclear to me if it possible to get compensation for damages by suing the manufacturer for copyright infringement. The amount of damages can be far from zero, for example, if the code is being dual-licensed: the non-GPL option can be a proprietary license, thus violating the GPL can be a way to the same software cheaper (no need to pay money nor respect freedom).

In most cases, there is no need to terminate the license. Violators can be sued for their actual breach of copyright. Terminating their license additionally makes it possible to sue them for legitimate use of the work. It doesn't make any sense in the context of the GPL, which is supposed to be about empowering users. Termination clauses are about putting the power firmly in the hands of the licensors, not the users--they're a staple of EULAs. The only case in which termination makes sense, IMHO, is for a party that files lawsuits (e.g. patent lawsuits) in breach of the GPL.

Comment 3235: What exactly are the 30 days for?

So what happens if an offender goes back into compliance with the license after the 30 days? Is the license reinstated only if permission is expressly granted by the copyright holder(s)?

/ Is the license reinstated only if permission is expressly granted by the copyright holder(s)? /

Yes, they have full power over this. In order to reinstate your license, they could demand large sums of cash, your car keys, that you submit to lengthy and vigerous sessions of anal sex, etc. Entirely up to them.

Comment 3236: 2 Critical problems with this section

It is one thing to require preservation of the Appropriate legal notices feature in existing interactive user interfaces.

It is entirely different to compel users to include such a feature in any newly created interactive interfaces. This is far worse than the equivalent clause in GPL v2. People WILL ignore this requirement, and assume it merely mandates preservation of the feature in existing interfaces. This is a *critical* problem with the license IMHO.

Further there is no exception for interactive user interfaces where it is *impossible* to meet the definition of "displays 'Appropriate Legal Notices'". Cases like an audio-only interactive interface could not possibly include "a convenient and prominently *visible* feature", and might not be able to tell "the user [..] how to view a copy of this License". It might not be possible to tell the user how to view the licence. Especially if the device has no means of visual output, in which case there is no way to "view" the licence. The program could perhaps give instructions on how to request to have the licence read to them, but that is not what is required.

That is also a critical flaw in the license that could be avoided by not mandating the inclusion of this feature in newly created interactive interfaces.

Most people don't read the source code. If a new user interface is created which doesn't display notices, users won't know their rights.

Comment 3237: Please remove this text

A patent troll, who is in the business of doing nothing but collecting fees for their patent portfolio, is not in the business of distributing software. A patent troll would have an effective defense against this clause. The phrase seems to weaken this part of the license. The clause should apply to everyone regardless of what business they are in.

Comment 3238: replacement

Recommend replacing "Ty Coon" with "Big Cheese"

Comment 3239: sexual connotations

Using a sexual position as a version number kind of detracts from the seriousness of the license... How old are we again?

I didn't even notice this until I read the comment. I don't think it would usually be an issue, but just to avoid any problems or comments, I would change it to any other number. I hereby propose 42.

I agree to mjuarez, but is it sure that the number 42 has any special meaning than universal concept in the other significant cultures?

The only symbolism I know of for 42 is its mention in the "The Hitchhiker's Guide the the Universe"...but probably has much more!

The inappropriate sexual reference makes the GPL into a joke. The International Declaration of Human Rights does not contain inappropriate sexual references. The Declaration of Independence of the United States of America does not contain inappropriate sexual references. Documents that are meant to be taken seriously do not contain inappropriate sexual references. By putting a sexual reference into the GPL, the FSF is telling the world that the GPL is a joke.

Much as I admire and miss Douglas Adams, it would probably be better to avoid 42 as well, because it would still be meant humorously. It would probably be better to be soberly business-like in the license language and avoid any appearance, to people considering adopting OSS, that we're treating the license as a joke.

Comment 3240: Contradictory

This section mandates that all UIs have notices, then goes on to talk about UIs that don't have notices! Apart from being far too restrictive, it doesn't even make sense.

It does need rewording. Perhaps something like changing "however, if" to "unless", and inserting "in which case".

Comment 3241: Very hard to parse

I had to read this several times, and I'm still not 100% sure what it says.

The "but which is not part of that major component" is the most confusing portion. I don't know what "part" means in this context.

Think of an implementation of the C standard library. It might commonly be packaged with compilers or kernels (compilers and kernels are Major Components), but isn't itself part of a compiler or kernel.

Comment 3242: Difficult to understand

This is the other passage I've been struggling with--I've read it several times, and it's still not clear how many different parties are involved, and what their relationships might be.

It's meant to assure compatiblity with the Apache license. Don't worry if you don't understand it. Neither did the FSF at first; that's why they took so long to add it. If I understand the rationale, it basically means that if:

1. Alan writes a program.

2. Bob redistributes it to Charlie.

3. Bob offers Charlie a warranty and says Bob will assume liability with for any problems Charlie has with the program.

Then:

4. Bob must ensure that if this somehow gets Alan sued, Bob protects Alan.

The prepositional phrases do get confusing. The main line of it is: "requiring indemnification of licensors ... by anyone who conveys ... for any liability...".

Comment 3243: Excellent

This whole paragraph (and the surrounding section) are marvelously clear. Very nice.

Comment 3244: Trailing slash required

As noted at the top of the document as a comment on the first web address, this needs a trailing slash. It has been noted by multiple people in the last draft. I don't see why it hasn't been altered. Was there not enough time for copyediting? An explanation for why a trailing slash is needed is in the first comment, but it suffices to mention here that this address is referring to a directory, in which case a trailing slash is necessary.

I suggest changing this to "http://www.gnu.org/licenses/" (sans quotes).

How do you know this address is referring to a directory? Even leaving aside future implementations of the website, it may do in the current implementation, or it may not. Given mod_rewrite and similar, you have no way of telling how their underlying filesystem is structured.

http://www.gnu.org/licenses is an entirely valid URL. It requests the resource "licenses" at the top level of the website served on port 80 by the machine with DNS name www.gnu.org.

Comment 3245: Definition of "the most recent violation"?

In case someone has conveyed a binary in violation of GPLv3-draft4, does "the most recent violation" mean the moment when

(1) the conveying was done (possibly long time ago); or

(2) the user notices the violation and requests the Corresponding Source, and gets a reply "no source code for you" or no reply at all?

The conveying itself was presumably not the violation - the violation would be any point at which you failed to satisfy the requirements of the license. I agre it would be better to reword it more clearly. Perhaps "the most recent failure to satisfy a responsibility required by the license".

If a written offer was not provided, then the conveying was the violation. If a written offer was provided, then there was no violation of the GPL, rather there was a violation of the written offer.

Comment 3246: Could it apply to customers

Seems like this could be used by someone to keep their own changes closed while giving the software with modifications to their customers, if the terms of the agreement are restrictive enough. If it is service provider, the running of this software would be under its control.

Would this allow, for example, fixing a bug in a GPLed fsck, and offering a service to customers (with filesystem corruptions that the unmodified fsck could not fix) where the customer gets the fixed fsck, is told to run it on the corrupted disk, and then delete the fsck given to him, and the customer is not allowed redistribution under GPL of the modified version?

This clause really wouldn't allow your scenario, because the peceding sentence says "for the sole purpose of having them make modifications exclusively for you". If they're not modifying the code, this clause doesn't apply.

Comment 3247: illegal in Germany

In Germany contracts about future wealth are contra bones mores and thus impossible. I think the same will apply to other legislations.

Hmm. I've noted the GPL doesn't have the usual clause about "if any bit of this is unenforceable, only that tiny bit is removed". I wonder why that is?

Are you sure this is illegal in Germany? That seems ridiculous. I would compare agreeing not to sue someone for patent infringement for patents you might receive in the future with agreeing not to shoot someone with a gun you might receive in the future. You're not committing to giving them something tangible, what you're giving them is a committment not to take from them. If this really was illegal, wouldn't it be a problem for patent cross-licensing deals?

Comment 3248: systematic problem

why definitions here again?

It's explained in the rationale document. They moved it here because it is used only in this section.

Comment 3249: worldwide patent license?

There is no world-wide patent license. these licenses don't exist on the market place. The draftperson fails to understand the territorial principle of patent law.

It's saying that the author licenses you, under all patents that the author holds that affect the Program, wherever those patents may be held.

That is, it's not the patents that are worldwide, it's the license.

Yes, I think it does make sense, given the definition of "essential patent claims".

Comment 3250: proprietary?

doesn't it make the software proprietary?

No, the software is free, but the associated trademarks are not. It seems [to me] perfectly reasonable that you can't claim your modified version is the version represented by the trademarks associated with the original version.

/ No, the software is free, but the associated trademarks are not. It seems [to me] perfectly reasonable that you can't claim your modified version is the version represented by the trademarks associated with the original version. /

It certainly is reasonable that you can't use trademarks on modified versions. However it is not clear to me that this clause applies only to modified versions.

Yes, it applies to unmodified works as well as modified ones.

However, the argument is basically the same. The trademarks are separate from the copyright and you don't need to free them in order to free the software.

I agree, though, that this clause should be extended slightly to say that "use of the trademarks, service marks, etc. by the unmodified software itself is allowed". That is, as recipient of a work you do not have the right to use trademarks commercially, but if the unmodified program describes itself using a trademark, that's allowed.

Comment 3251: is the novell/microsof deal being "grand-fathered" in?

Quote from (http://www.linux-watch.com/news/NS8853808815.html)

"At the same time, the controversial Novell-Microsoft patent deal has been "grandfathered" into the GPLv3. A new clause will let companies -- like Novell distribute GPLv3 software even if they have made such patent partnerships, so as long as the deal was made before March 28."

So it seems this decision in GPL3 isn't really subject to debate since it hasn't changed since draft 3 where there was strong opposition to it.... Can't this open a loophole where someone that entered a deal after March 28th tries to (and is able to) claim that they have entered it prior to March 28th?

Comment 3252: terminology program

above software was used,

Comment 3253: "rights"

I wonder what the difference of "rights" and "freedoms" is.

Comment 3254: Which "you"?

Let's take, as an example, company A, which wants to make something that violates GPLv3. It won't do anything of the sort itself, but founds a company A1, which will perpetuate the act. When A1 eventually receives the notice of violation, they'll apologize and say that they won't do that again.

What would prevent A from shutting down A1 and starting a company A2 that would start all over with the license violations? Could A use this to violate the license in perpetuity? Won't 30 days be enough to transmit the first time violator status smoothly over and over again?

I checked the draft 3 and similar concerns were voiced in its comments too. See, for example, 2466.

Comment 3255: Recognized by who?

It might be better to leave a pointer to a list of standards bodies FSF recognizes.

Comment 3256: Collect all definitions here, please

It is nice to have definitions in their own section, but there are some definitions still outside this section. They should be moved here.

Definitions which apply across sections are in this section; definitions which apply only in particular sections are in that specific section. See the draft 4 rationale document relating to the moving of one definition from here into clause 11.

Comment 3257: inclusion into section 7?

is there any reason why the terms in which the GNU AGPL differs from the GNU GPL, are not included as a possible additional restriction in section 7?

That's the way it was done in draft 2. Read the comments in there and the rationale document for draft 3.

Comment 3258: typo

typo: missing space, should read "Foundation, Inc.,"

Comment 3259: Prevent VS Cure.

Prevent VS cure? How long will take to cover most open source softwares with GPLv3? That will take a long time... So we should prevent abuses NOW. It should be nice to remove this escape. It also creates favouritism for one big company (I don't want to quote it).

Comment 3260: Nasty consequences for freedom to tinker

I haven't been following this, so I'm sorry if this repeats some previous comment.

Under this, a producer who wants to sell locked-down hardware to the commercial and industrial markets needs to be sure to avoid having it classified as a "User Product". The obvious way to do that is to refuse to sell the product to anybody who even *might* be a consumer, and perhaps to require that recipients agree not to sell it on to consumers.

That will create a whole class of products that you and I can't buy *at all*, let alone buy and get the benefit of anti-tivoization. In fact, we may have trouble buying the products even for commercial purposes, if we can't produce the right sorts of evidence that that's what we intend.

I've bought, for personal use, a lot of things that sellers meant for the commercial and industrial markets. I'd like to be able to keep doing that, and I'd like to be able to get GPL software in those devices. I also think it's socially unhealthy to create a distinction between "people" and "producers". Individuals should be encouraged to create for themselves, and sometimes that means using industrial tools. Freedom to tinker is an important value.

In any case, to my not-a-lawyer eyes, the entire "User Products" hack seems fraught with complexity, uncertainty and litigation danger.

I understand that the reason for this is that businesses sometimes *want* to buy locked-down hardware, for internal security reasons or for regulatory reasons. Why not address that by replacing the entire "User Products" structure with a voluntary exemption structure?

Suppose you say that the obligation to convey Installation Information does not apply if, at the time that the product was originally conveyed, the recipient voluntarily requested that no Installation Information be provided for that product, with no other part of the transaction being conditioned on that request. In other words, let me specifically order a "locked down" device without keys, or a "not locked down" device with keys, and require that the producer be prepared to provide either one.

That might require that the keys for each unit be different, but that's good security practice anyway.

Comment 3261: Specifically delimit where does the short notice start and end.

Right now, this line does not specify clearly where does the short notice end. Maybe use an End-of-Notice marker, or something like --START GPLv3 NOTICE-- and --END GPLv3 NOTICE--.

Comment 3262: Vice is ambiguous

I assume Vice here is referring to a company/corporation. However, it would be more clear if the text said something like "Vice Inc." or "Vice Corporation"

The entire section should be cleaned up to remove the pathetic attempts at humor.

Vice isn't a corporation; the analogy is "President of Next-Generation Networks" or "President of Marketing" -> "President of Vice".

There's no point putting "Vice Inc." or "Vice Corporation" since the word "Vice" was chosen because of the double meaning--"evil", and "deputy" as in "Vice President". However the license should avoid double meanings for the sake of people who speak English as a second language, and so that translations can be equivilant to the original.

Comment 3263: What about positive powers?

In addition to waiving powers to *forbid* circumvention, shouldn't this bind the licensee to exercise any powers it may have to actively *permit* circumvention?

Comment 3264: Why is this needed here?

The license accompanies a specific software, which in this case is assumed to be Gnomovision. Why is this clarification needed here, and how long should it be? "Making passes at compilers" is too broad, IMHO. Also, anything that is product specific, and trying to be unambiguous, would have to be too long, and if anything, would need it's own section.

Comment 3265: Narrow this to T&C

Although I think it would be better to allow people to make their own modified licenses, so long as they don't identify it as the GPL, if you won't do that, at least narrow the scope of the "changing it is not allowed" to just the Terms and Conditions, so people who prefer their licenses business-like can drop the polemical preamble and the embarrassing juvenile humor of the "How to Apply" section. By definition, those sections have no legal value, anyway, so let us excise them if we choose.

How would changing this phrase be useful? Good points have been made, but most of them are theoretical and offer no examples of how to better protect the four freedoms. The ability to change the document is also the ability to restrict the very thing it's trying to protect. The license has to be an enforceable document in the courts, and in that sense must play by the copyright system's rules. Being able to modify it would make it more of a template for a license than an actual license. This seems to be more of a philosophical (though valid) quibble than an outright flaw.

How would changing this phrase be useful? Good points have been made, but most of them are theoretical and offer no examples of how to better protect the four freedoms. The ability to change the document is also the ability to restrict the very thing it's trying to protect. The license has to be an enforceable document in the courts, and in that sense must play by the copyright system's rules. Being able to modify it would make it more of a template for a license than an actual license. This seems to be more of a philosophical (though valid) quibble than an outright flaw.

Note that my suggestion in this comment did not allow changing the license's legal parts, just the non-binding wrappers. That arguably makes the license usable by more people, since some people do object to the polemic and the unfunny jokes.

However, as noted, I do think it would be better to allow modifying the license, because the FSF is no more perfect than the rest and is possible someone will come up with a better (even from the point of view of the four freedoms) license based on the GPL. It also seems somewhere between inappropriate and hypocritical for the FSF to say "freedom to modify software to get just what you want is good, but freedom to modify licenses so you get just what you want is not."

If removal the Preamble and How to Apply sections would be allowed, the modified license would no longer be *the* GPL. Thus, the new license would by definition be incompatible with the GPL. That would increase license proliferation, which we obviously don't want.

I find it better that if someone wants to make a new license based on the GPL, he or she has to ask the FSF for permission.

This is simply not true. "Compatibility" under the GPL means allowing distribution, etc., under the same terms as the GPL. Removing hte Preamble and "How to Apply" have NO impact on the terms and conditions and would be completely compatible with the GPL. It's not proliferation in any sense.

Sorry sepreece, I misread your earlier comment a bit.

I agree ENTIRELY with sepreece. Removing the Preample and "How to Apply" will not make it incompatible.

I also agree that the license itself should be free. It's the ethics behind the GPL in the first place. It should allow modifications, so long as they do not falsely claim to be the GPL, which could cause issues with people not knowing which license they are following.

I like the idea of the license being free though it possibly should be forbidden to use the the name of the GPL for a derivative work so the GPL and the derivative work cannot be mixed up.

Allowing specifically to drop the Preamble and the "How to Apply" section might make sense legally but I think those parts are actually important for people who do not know yet what the GPL is. The preamble covers the meaning and thought behind free software and if just one person reads the GPL and learns about free software that is a Good Thing.

Let me add that the sentence at hand says one cannot modify the License *document*, not the License. That seems right, as of course the License is an "entity object" which can't be modified by simply altering a copy (say e.g. you alter a tape registration of a song: you haven't modified "the song" :-)). However I still see two problems with the wording: - it should use the term "modify", and point to paragraph 0 of Terms and Conditions where that term is defined - it should specify what happens when e.g. the FSF address changes: I'd see something like "[Note: the Free Software Foundation will make updated copies available in case their address, or other data which don't affect the terms and conditions of the License, might change]"

What do you think?

Comment 3266: Novel, Samsung, and now Xandros???

After the Novel March 28 exception date Microsoft has signed the same deal with Samsung and now it has just been reported that Xandros has also signed a deal with them. Apparently it looks like they don’t see much of a limitation in GPL3 or a lot of people are not taking it seriously. How many other deals are out there like that which we have yet to hear about? Why should we allow any exception for Novel? There should be no exceptions.

Why bet the GPL3 farm on the possibility that the Novel/Microsoft deal will subject Microsoft to GPL3 on the notion that (Distributing coupons = distributing the code) are you crazy? They will probably argue something clever live those are coupons for providing the technical support contract of Linux and not the code. I think there is just too much wiggle room and they will use it.

Comment 3267: Any FSF representative care to comment on this?

I'm sorry, but I have to ask if someone who represents the FSF or can speak on its behalf care to explain why exactly this date was chosen? Why exactly should Novell be diferent from anyone else? Why is this possible loop-hole/manouvering-space even be in GPL3?

Since this is the last draft its the last chance to change anything...please listen to the people who will eventually end up using the license itself!

Read the Rationale document (http://gplv3.fsf.org/rationale), available on the license website. It specifically explains why the date was left in. The Rational document is presumably representative of the FSF's thinking on the topic.

I read the Rationale document, but I still can't agree with the reasons. Here is why:

1-The Free Software Foundation says that Microsoft is basically screwed either way because of the coupons it acquired from Novell. However, there has already been talk about how Microsoft could wiggle out of this since the coupons are for SUSE support, not for direct SUSE distribution. And we know how long court proceedings can take...and MS has *much* deeper pockets than SCO

2-There is a greater benefit to let Novell continue to distribute GPL3 code and continue its pact with MS than to deny it to do so. So then what makes Novell better than Xandros, that has now also jumped into bed with Microsoft? Novell betrayed the community by literally endorsing Microsoft's patent portfolio, why should they (of all people) get a special clause in GPL3 that lets them continue with their pact?

I read the Rationale document, but I still can't agree with the reasons. Here is why:

1-The Free Software Foundation says that Microsoft is basically screwed either way because of the coupons it acquired from Novell. However, there has already been talk about how Microsoft could wiggle out of this since the coupons are for SUSE support, not for direct SUSE distribution. And we know how long court proceedings can take...and MS has *much* deeper pockets than SCO

2-There is a greater benefit to let Novell continue to distribute GPL3 code and continue its pact with MS than to deny it to do so. So then what makes Novell better than Xandros, that has now also jumped into bed with Microsoft? Novell betrayed the community by literally endorsing Microsoft's patent portfolio, why should they (of all people) get a special clause in GPL3 that lets them continue with their pact?

Comment 3268: Potential violation leakage from Novell?

I am concerned about this. I'm afraid it took a while for me to realise this might be possible.

http://www.cbronline.com/article_news.asp?guid=FB198B0A-7B41-4DF1-848B-B4D7C534E6C1

Now, as was reported at the time, when source code for Windows 2000 leaked, many in the FOSS community refused to look at it because of the possibility that they might accidentally copy code. This is the process many of us have experienced where you think you have created an original idea only to find that you actually saw, heard or read it somewhere else and it is not your idea at all.

What is to prevent this occurring at Novell or Xandros? What if this occurs with a GPL licensed project? If a programmer at Novell should have access to Microsoft source code and accidentally applies a patented 'method' in a GPL project at a later date, what protection does anyone who does not work for Novell have?

It is Novell's position that they sought this patent protection for their _paying_ customers in case something like this should happen.

What does this mean for everyone else? Unless I have missed something, the only way to guarantee protection from this possibility is to not accept any code or binaries from Novell or Xandros (assuming Xandros have the same access).

Does this make the code unredistributable and, if so, is this not somehow a breach of the GPLv2 (conferring same rights to recipients to redistribute)?

Maybe I've missed something, I hope so. For the time being this does not affect me but I'd hate to see anyone affected by software patents suffer due to the wilful circumvention of the GPLv2.

It is impossible to guarantee that such a leakage of information will not occur and unless Novell and Xandros can assure some kind of protection from legal threat I will be very wary of any code which comes from within their walls.

Comment 3269: Contradiction

"Additional permissions that are applicable to the entire Program shall be treated as though they were included in this License," with: "When you convey a copy of a covered work, you may at your option remove any additional permissions from that copy,"

-removable "requiring and prohibiting"?? So: if i receive something with AP, i remove them "at my option" and have to add my copyright notice even without having modified the program itself, just cos i 'modified some license conditions'? I think this whole section comes as a list of suggestions for some authors to be more confident, innecesary cos: 1- d. there are moral rights on copyright law impeding this!. Duplicates and gap on freedom 0. 2- a. + reasonable(add full copyright law +full us constituiton + ... re: Apropiate legal notices if the nature of the program could have some liability implications cos it's solely use, not meaning the production made with it) legal notices b. + .e + .f = could go(or already go) in another general uncategorized section as a explain, so duplicates here. 3- atribution on b.= it doesn't come with a definition, so leaves a dnagerous gap which could be used malevolentemente to, for example: include a undeletable 10 (macromedia)flash presentation of the author attributes before running the program, etc...

Comment 3270: compatibilty

what about the way ccbysa 3.0 deals with this at its section 1c? (it does include an (empty now) url where to see the compatible ones, which would give some freedom for the acceptance's scheduling for adding new ones)... Could it also include compatibilty for linking with the 'the future gnugplv4' for being able to link v3 programs with v4 ones which don't choose at v3 the 'or any later' ? cos of : "There is no way to make them compatible(v2 & v3). We could add a GPLv2-compatibility clause to GPLv3, but it wouldn't do the job, because GPLv2 would need a similar clause."

Comment 3272: Enclosing URL in something

I think the URL should be enclosed in angle brackets (or non-URL character separates URL and period,) to distinguish "http://www.gnu.org/licenses" and "http://www.gnu.org/licenses.". It is possible to decide which URL is the actual URL (since we all know that all valid English sentence is ended with period, and period-ending URL is so rare) but it can confuse the people without the background information about URL or the text preprocessor program.

Comment 3274: An unfair concession

It is unethical and unjust to allow Novell and others in similar patent agreements with M$ to continue their business as usual without any penalty. Which section of GPL3 is going to reward businesses like Redhat which publicly opposed such deals? If we give this concession to Novell today, we will owe another cut-off date in GPL4 to other businesses that couldn't benefit from this date today.

Comment 3275: Possibly excessively visually centric

Some have suggested that instructions to display required information are excessively visually centric and fail to consider other means of information output when visual output is unavailable. I agree with the sentiment. although, it is not a significant issue for me at the present time. Some modest consideration should be given to addressing issues which may not be significant issues now but may become significant issues in future. The original definition for 'Appropriate Legal Notices' seems to only have problems for visual centricism before the last sentence. Some have suggested otherwise, that the last sentence itself is problematic in relation to visual centricism. However, criticisms of the last sentence were merely objecting to the very concept of a requirement from section 5 d to inform the user of necessary information in a manner in which most users are liable to actually notice the necessary information. I have answered those objections in that context. Original: "0. Definitions." ... " An interactive user interface displays "Appropriate Legal Notices" to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion." 1. SUGGESTED CHANGE. If a remedy to this issue is introduced, great care must be taken to avoid the risk that someone modifying the program might find cover in the language to downgrade required information output to a form least likely to be understood by most users when output more readily understood by most users is available. A related risk is that some less common output methods may not be properly supported or configured on a particular user's system. Unfortunately, the care needed to avoid those risks seems to need a new rather verbose definition for 'Presents' which is substituted for 'displays'. The required verbosity seems to leave any attempt to include an appropriate change completely within the definition of 'Appropriate Legal Notices' much to awkward even when compare with my suggested change with a definition inside a definition. I believe my suggestion is least awkward with regard to the issue of introducing a new definition or not. I imagine it may be improved provided due care is taken to avoid the cited risks. Note that my suggested definition of 'Presents' does not preclude using multiple output methods so that the blind and the deaf may both be informed even if their screen readers could not work adequately with the visual output provided. My suggested definition merely precludes substituting the most helpful output method for most people for a less helpful output method for most people when the more helpful output method most people is available to present the required information. 1.1. DIFF OF SUGGESTED CHANGE. In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d. [[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}} "0. Definitions." ... "{{ An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information.}} An interactive user interface [[displays]] {{Presents}} "Appropriate Legal Notices" to the extent that it includes a convenient and prominently [[visible]] {{apparent}} feature that (1) [[displays]] {{presents}} an appropriate copyright notice, and (2) [[tells]] {{informs}} the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to [[view]] {{examine}} a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion." 1.2. COMPLETED SUGGESTED CHANGE.. In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d. "0. Definitions." ... " An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information. An interactive user interface Presents "Appropriate Legal Notices" to the extent that it includes a convenient and prominently apparent feature that (1) presents an appropriate copyright notice, and (2) informs the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to examine a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion." 2. ATTENTION TO RISKS. As explained above, this issue is not significant for me currently even though I support the sentiment. I took the time to address the issue in detail not merely because the issue might be significant for me in future and may be significant for others now, but importantly because of my concern that great care must be taken to avoid the cited risks if any effort is taken to correct visual centricism in the license.

testing

line breaks

which failed

Comment 3276: rules for communication

assume there is a broadcast network with a protocol that requires users not to store contents (by protocol specification). could a receiver box then allow you to _run_ your modified copy, but prevent it from accessing the network?

Comment 3277: placeholders

in other parts of this section, placeholders are either marked with <> ("") or filled with example data ("Gnomovision version 69"). writing "year name of author" here is inconsistent with that and hard to read anyway.

Not only that, but the marked text shouldn't even be using place holders. It's part of an example, not a template.

Comment 3278: Are Microsoft's coupons really distribution of GPL'd software?

According to the rationale, this was kept because Microsoft distributing coupons is being interpreted as distribution of GPL'd software. If this interpretation of coupons is correct, keeping the date in GPLv3 might be a good idea.

If coupons are not interpreted to be distribution of GPL'd software, leaving this date in the final GPLv3 may backfire badly: Microsoft wouldn't be bound by GPL in any way, and thus not extending the patent "license" to everyone. At the same time Novell would be allowed to continue taking advantage of its deal with Microsoft thanks to this special date in GPLv3.

I hope that you are very sure that distribution of the coupons equals distribution of GPL'd software. To me it sounds like a safer bet to leave this date off of the final GPLv3.

Comment 3279: perhaps excessively visually centric

Attempted repost hoping line breaks will be fixed.

Some have suggested that instructions to display required information are excessively visually centric and fail to consider other means of information output when visual output is unavailable. I agree with the sentiment. although, it is not a significant issue for me at the present time. Some modest consideration should be given to addressing issues which may not be significant issues now but may become significant issues in future.

The original definition for 'Appropriate Legal Notices' seems to only have problems for visual centricism before the last sentence. Some have suggested otherwise, that the last sentence itself is problematic in relation to visual centricism. However, criticisms of the last sentence were merely objecting to the very concept of a requirement from section 5 d to inform the user of necessary information in a manner in which most users are liable to actually notice the necessary information. I have answered those objections in that context.

Original:

"0. Definitions."

...

" An interactive user interface displays "Appropriate Legal Notices" to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

1. SUGGESTED CHANGE.

If a remedy to this issue is introduced, great care must be taken to avoid the risk that someone modifying the program might find cover in the language to downgrade required information output to a form least likely to be understood by most users when output more readily understood by most users is available. A related risk is that some less common output methods may not be properly supported or configured on a particular user's system.

Unfortunately, the care needed to avoid those risks seems to need a new rather verbose definition for 'Presents' which is substituted for 'displays'. The required verbosity seems to leave any attempt to include an appropriate change completely within the definition of 'Appropriate Legal Notices' much to awkward even when compare with my suggested change with a definition inside a definition. I believe my suggestion is least awkward with regard to the issue of introducing a new definition or not. I imagine it may be improved provided due care is taken to avoid the cited risks.

Note that my suggested definition of 'Presents' does not preclude using multiple output methods so that the blind and the deaf may both be informed even if their screen readers could not work adequately with the visual output provided. My suggested definition merely precludes substituting the most helpful output method for most people for a less helpful output method for most people when the more helpful output method most people is available to present the required information.

1.1. DIFF OF SUGGESTED CHANGE.

In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d.

[[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}}

"0. Definitions."

...

"{{ An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information.}} An interactive user interface [[displays]] {{Presents}} "Appropriate Legal Notices" to the extent that it includes a convenient and prominently [[visible]] {{apparent}} feature that (1) [[displays]] {{presents}} an appropriate copyright notice, and (2) [[tells]] {{informs}} the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to [[view]] {{examine}} a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

1.2. COMPLETED SUGGESTED CHANGE..

In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d.

"0. Definitions."

...

" An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information. An interactive user interface Presents "Appropriate Legal Notices" to the extent that it includes a convenient and prominently apparent feature that (1) presents an appropriate copyright notice, and (2) informs the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to examine a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

2. ATTENTION TO RISKS.

As explained above, this issue is not significant for me currently even though I support the sentiment. I took the time to address the issue in detail not merely because the issue might be significant for me in future and may be significant for others now, but importantly because of my concern that great care must be taken to avoid the cited risks if any effort is taken to correct visual centricism in the license.

Line breaks are better with my repost attempt. However, a blank line just before the definition of 'Appropriate Legal Notices' is missing. The following text tests if closing the quotation marks at the end of the paragraph will correct the problem.

1.1. DIFF OF SUGGESTED CHANGE.

In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d.

[[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}}

"0. Definitions."

...

"{{ An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information.}}"

" An interactive user interface [[displays]] {{Presents}} "Appropriate Legal Notices" to the extent that it includes a convenient and prominently [[visible]] {{apparent}} feature that (1) [[displays]] {{presents}} an appropriate copyright notice, and (2) [[tells]] {{informs}} the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to [[view]] {{examine}} a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

Comment 3280: Use International English

We are told that the GPL is an international document, but it's written in US English. English is an official language in about 70 countries, and is fluently spoken by about 1,000 million people (about 2,000 million people have some knowledge of English), of which only 1 country including about 260 million English speakers uses US English spelling (i.e. the US). Thus the GPL is written to cater specifically for this 1/4 of English speakers. Show us that the GPL is an international document--write it in international English.

Comment 3281: Non-commercial distributors responsibility to provide source should be non-transferable

As it stands, someone receiving a GPL work can redistribute non-commercially, and transfer the responsibility to provide source back to the person that supplied them, even if that person was also conveying non-commercially.

Make it "if you received the object code /commercially/ and with such an offer".

Comment 3282: Maybe excessively visually centric

Reposted comment to fix missing line breaks breaks for the last time, I hope.

Some have suggested that instructions to display required information are excessively visually centric and fail to consider other means of information output when visual output is unavailable. I agree with the sentiment. although, it is not a significant issue for me at the present time. Some modest consideration should be given to addressing issues which may not be significant issues now but may become significant issues in future.

The original definition for 'Appropriate Legal Notices' seems to only have problems for visual centricism before the last sentence. Some have suggested otherwise, that the last sentence itself is problematic in relation to visual centricism. However, criticisms of the last sentence were merely objecting to the very concept of a requirement from section 5 d to inform the user of necessary information in a manner in which most users are liable to actually notice the necessary information. I have answered those objections in that context.

Original:

"0. Definitions."

...

" An interactive user interface displays "Appropriate Legal Notices" to the extent that it includes a convenient and prominently visible feature that (1) displays an appropriate copyright notice, and (2) tells the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to view a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

1. SUGGESTED CHANGE.

If a remedy to this issue is introduced, great care must be taken to avoid the risk that someone modifying the program might find cover in the language to downgrade required information output to a form least likely to be understood by most users when output more readily understood by most users is available. A related risk is that some less common output methods may not be properly supported or configured on a particular user's system.

Unfortunately, the care needed to avoid those risks seems to need a new rather verbose definition for 'Presents' which is substituted for 'displays'. The required verbosity seems to leave any attempt to include an appropriate change completely within the definition of 'Appropriate Legal Notices' much to awkward even when compare with my suggested change with a definition inside a definition. I believe my suggestion is least awkward with regard to the issue of introducing a new definition or not. I imagine it may be improved provided due care is taken to avoid the cited risks.

Note that my suggested definition of 'Presents' does not preclude using multiple output methods so that the blind and the deaf may both be informed even if their screen readers could not work adequately with the visual output provided. My suggested definition merely precludes substituting the most helpful output method for most people for a less helpful output method for most people when the more helpful output method most people is available to present the required information.

1.1. DIFF OF SUGGESTED CHANGE.

In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d.

[[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}}

"0. Definitions."

...

"{{ An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information.}}"

" An interactive user interface [[displays]] {{Presents}} "Appropriate Legal Notices" to the extent that it includes a convenient and prominently [[visible]] {{apparent}} feature that (1) [[displays]] {{presents}} an appropriate copyright notice, and (2) [[tells]] {{informs}} the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to [[view]] {{examine}} a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

1.2. COMPLETED SUGGESTED CHANGE..

In addition to the changes shown below, minimal compatibility changes would be needed elsewhere in the license such as changing 'display' to 'Present' in section 5 d.

"0. Definitions."

...

" An interactive user interface "Presents" when the user interface at a minimum: (1) has display output which displays required information; or (2) has printer output which prints required information, but has no display able to able to display textual information; or (3) has sound output which speaks required information, but has neither display nor printer output able to display or print textual information; or (4) has tactile output which impresses required information, but has neither display, printer, nor sound output able to display, print, or speak textual information."

" An interactive user interface Presents "Appropriate Legal Notices" to the extent that it includes a convenient and prominently apparent feature that (1) presents an appropriate copyright notice, and (2) informs the user that there is no warranty for the work (except to the extent that warranties are provided), that licensees may convey the work under this License, and how to examine a copy of this License. If the interface presents a list of user commands or options, such as a menu, a prominent item in the list meets this criterion."

2. ATTENTION TO RISKS.

As explained above, this issue is not significant for me currently even though I support the sentiment. I took the time to address the issue in detail not merely because the issue might be significant for me in future and may be significant for others now, but importantly because of my concern that great care must be taken to avoid the cited risks if any effort is taken to correct visual centricism in the license.

Comment 3283: Allow distribution of source at twice cost

The requirement to provide source code at cost is unrealistic. It is difficult enough to convince stores to stock GPL software that earns them a pittance, drastically undercutting the income they make off commercial software--although a few do. Requiring them to provide source code at cost is just shooting Free Software in the foot. If anyone ever enforces this requirement, then Free Software will lose these distribution channels and the credibility they provide. Free Software is not supposed to be anti-business, but this requirement is. Twice cost is reasonable, really.

Stores that stock GPL software are not licensees. They have no obligation to provide source code. Moreover, they can price the software they sell at any amount they fancy.

In general source is provided by licensors free of charge, because the organisational hassle of charging for it is not worthwhile (unless you already have a retail facility that can handle burning DVDs on demand to despatch). Either way, a profit is not warranted. The profit was obtained when the software was sold.

Either I've completely missed your point, or worse, you've missed an important aspect of the GPL. Apologies if the former.

/ Stores that stock GPL software are not licensees. They have no obligation to provide source code. /

This is true for pseudo-free software like that from Novell and Xandros--this comes shrink-wrapped from a factory. True Free Software though, generally does not. OpenOffice.org provide ISOs for distributors to burn their own CD from. Until recently they didn't even do that.

Here in New Zealand, Dick Smith Electronics sells home-brand OpenOffice.org CDs. Their URL is "http://search.dse.co.nz/", you can search for "DSE OpenOffice.org Suite 2.0 OS CD" or "XS0793". The product image can be magnified, and as you can see, it is a DSE branded product.

/ Moreover, they can price the software they sell at any amount they fancy. /

Yes, but realistically, if they sell it at all, they have to price it high enough that they don't make a loss on each copy, but low enough that people actually buy it. I suspect this is a fairly small window, and I doubt DSE would lose much by pulling OpenOffice.org from the shelves.

From a Free Software advocate's point of view, however, having DSE stock OpenOffice.org makes things easier, and more importantly I think, lends credibility. Insteading of trying to tell someone they don't need a shrink-wrapped product, 'cause I've got one I can copy onto a CD-R for them no-worries-mate (which, believe it or not, doesn't always go down that well), I can tell them they can get a good inexpensive product from DSE.

It would be much more of a loss for OpenOffice.org, than for DSE, IMHO, if DSE were to pull OpenOffice.org from the shelves.

/ In general source is provided by licensors free of charge, because the organisational hassle of charging for it is not worthwhile (unless you already have a retail facility that can handle burning DVDs on demand to despatch). Either way, a profit is not warranted. The profit was obtained when the software was sold. /

Again, fine for pseudo-free products, but distributors of true Free Software don't make a big profit.

/ Either I've completely missed your point, or worse, you've missed an important aspect of the GPL. Apologies if the former. /

I thought the GPL was about freedom, not price. If I'm wrong about that, then someone really has missed the point.

Yes, DSE is a licensee/licensor, so they are in the source chain (if they don't include all the source on their CDs), but then they may have reseller stockists (who wouldn't be in the source chain).

I'm interested to hear of this 'True Free Software'. Will there be another FSF license for this? Making 3 variations: GPL, AGPL, and TGPL?

Frankly, I'd be happy to allow binaries to be sold separately from source, and no obligation to disclose source, subject to all remaining copyleft. So in principle, I'd say anyone can charge what they like for source code. However, I was responding to your comment from the perspective of the GPL as I understand it.

Yes, of course, GPL is supposed to be about freedom not price, but whilst the non-obfuscation requirement remains - the obligation to provide source code (at no more than cost) - then that's how source is necessarily non-profit.

I actually don't think the FSF would have a big problem with saying "twice the reasonable cost" - as you say, it's about freedom, not cost.

However, I also don't see your argument as very convincing. If a distributor doesn't want to be bothered with providing the course code at cost, it can simply include the source code with the product and have no further obligation at all.

Note that "at cost" means what it says - you're allowed to recover what it costs you to provide the service, which could include somebody's time making the copy, the media, packing it, and shipping it; it's not just the media cost. However, it has to be reasonable (you can't say "well, the only way I could package it was in this titanium box that costs $15K").

The software purchaser has already purchased the source code at the time they purchased the software (the GPL binds the sale of source with the sale of software even if supplied separately).

Now, if the purchaser is enjoying a cheaper price (packaging/media/bandwidth) because the vendor saves on the cost of including the source, then naturally, they're entitled to recover this cost saving from the purchaser should they choose to require delivery of the source.

To enable a profit on separate provision of the source then creates an incentive for vendors to provide the source separately (more profit).

Somehow I think the last thing the FSF wants to do is to encourage vendors not to include source code with the software.

This birdie ain't going to fly. :-\

Thanks for all the feedback. I still don't agree, but it has helped a lot to sort out my thoughts. I've posted a new comment--Comment 3364: Allow providing source at market rate rather than cost

Comment 3284: Use of the work on a public server should not be unlimited

Violators using the unmodified work on public servers get away scott-free. Although those wishing to modify or convey works may be punished for violations and/or have their licenses terminated, those using the unmodified work on a public server are subject to nothing--they don't even have to accept the license. They can sue whoever they like over patents or whatever, and keep using the work themselves on a public server, and there's not a thing anyone can do about it.

Yes, and that's the way it should be. The unmodified copy is yours to use as you see fit.

/ Yes, and that's the way it should be. The unmodified copy is yours to use as you see fit. /

Even if you deny others the same right?

And why is it specifically use of the unmodified work that should be protected? Why not protect conveying too? And why not protect use of the modified work?

It seems to me that conveying is not protected (and ought not be) because it effects others, but use on a public server also effects others. Private use of a modified work doesn't though, and if anything, it is this that ought to be protected, IMHO.

I think perhaps use of the unmodified work is protected because it is the right usually allowed by propritary 'freeware' like Adobe Acrobat and Macromedia/Adobe Flash. It is not a protected right for these programs, it is the only right. Their EULA has a termination clause much like the GPL one, but it applies to this right. So the core right for the GPL is the 'freeware' right, perhaps by convention.

But the GPL is not designed with the same goals as 'freeware', it is entirely different, and this unmodified/modified distinction does not seem appropriate to me, I think a private/public distinction would make much more sense for the GPL.

Because "using" the work does not require the copyright owner's permission. The right to use is what you get when you acquire the work. The FSF has sensibly resisted attempts to add use restrictions.

And, of course, nothing is being denied to anyone - all others have exactly the same right to use the softwae that they would have in the absence of its use on a public server.

[Note: The US Copyright Act does reserve the right of public performance to the copyright holder, but only for a list of specific kinds of works, which does not include computer programs (a class defined in the Act and treated differently than other kinds of works in various places). Some people think that "audiovisual works", (which is in the list but is defined in a way that seems to exclude most computer programs), could be read to include at least some kinds of computer program; others disagree. The FOSS community has generally been opposed to that kind of expansive reading of the Copyright Act.]

I think this discussion is too abstract, and I could better explain what I mean with an example.

Say there are a couple of book-sellers in competition, aggressive company A and benign company B. Someone writes a killer GPL integrated application for on-line shopping, stock management, ordering from suppliers, accounting, etc., and both companies switch to it. The GPL application infringes some patent that A holds (say one-click shopping, a patent they got for their previous in-house software), and at some point, A's legal department realises that they can prevent competition by suing B for infringing on the patent, so they sue B. (Not a great example, but hopefully you get the idea.)

Now there are two possibilities. Either A is using the GPL work unmodified, or they are using a modified work (perhaps tweaked and recompiled for Solaris on Sparc, and hacked to work with a database backend their sys-admin likes).

If A is using the GPL work unmodified, they get away scott-free. If they are using a modified work, then they can't take any more orders (they have to take down the public on-line shopping site), and more than that, they can't even fill the orders they currently have, since they can't use the work privately either. They can't complete their tax returns, they can't even wipe their arses (the GPL software also controls access to the sanitation block).

To me, this enormous disparity seems bizarre and manifestly unfair. What is the moral difference between using the unmodified work, and using a modified work?

/ Because "using" the work does not require the copyright owner's permission. The right to use is what you get when you acquire the work. The FSF has sensibly resisted attempts to add use restrictions. /

I was given to understand that since copying the work requires permission, and since you cannot use computer software without it being copied (from hard-drive to memory, to swap-file, back to memory, to cache, to CPU), therefore use of computer software requires copyright permission. Besides, don't copyright licenses often include use restrictions? Am I wrong, or are these licenses illegal? Surely this is not really a legal issue?

And the GPL does include use restrictions on modified software--so if allowing unlimited use were a moral issue rather than a legal one, then why limit use of the modified software? This doesn't make sense either.

/ And, of course, nothing is being denied to anyone - all others have exactly the same right to use the software that they would have in the absence of its use on a public server. /

But the same could be said of someone conveying the work. If you are conveying the work while suing others, then your conveying is not preventing others from conveying, but it is providing both the incentive and the means to do so--you can make money by being able to convey while others can't, and you can put some of that money into further lawsuits. It seems to me that this applies for use on a public server just as well as for conveying.

Comment 3285: Termination clause far too harsh

The termination clause almost looks like it is from a commercial EULA (literally). Under the GPLv2, one tiny mistake meant eternal termination. Under the GPLv3, two tiny mistakes could mean eternal termination. It's better, but still draconian as hell. People are not infalliable, they make mistakes. The licensors are supposed to be allowing users freedom, yet they have you by the balls if you accidentally put a foot wrong. There is no need for this. You don't need to terminate all of a users rights in order to punish them for a breach of the license. Copyright law is harsh enough as it is--more than enough, IMHO. Termination should be reserved for people who file lawsuits against the GPL.

Comment 3286: Rewrite the SFDL as a set of permissive exceptions to the GPL

The GNU licenses prohibit format shifting. The GPL covers programs and the SFDL covers works for human appreciation. It is not legal to incorporate material from an SFDL tech manual into GPL literate programming comments, or from an SFDL story into a GPL game, or from SFDL teaching material into GPL interactive teaching software. The GNU licenses prohibit these kinds of use.

Comment 3287: military imagery

Small point: "we stand ready" needlessly evokes military imagery, IMO. Would "we are ready" be as good?

It's also a secondary or metaphorical meaning of "stand". It would be clearer, especially for second language speakers, to use words only for their primary meanings.

Comment 3288: Easy loophole?

To my understanding, this requires, that the amount of payment depends on how many copies of the software gets distributed. What if Novell and Xandros paid (or got paid) a one-time fee to get Microsoft to promise to not sue paying customers of the respective companies? Sounds that this paragraph wouldn't apply then, because one can argue that one-time fee isn't based on activity of conveying.

Unfortunately removing the highlighted part of the sentence could have other problems: Microsoft could then give anyone buying a copy of a Microsoft product a promise to not sue anyone who gets GPL'd software from the buyer of the Microsoft product. Thus, anyone buying Windows would be unable to convey software under the GPLv3.

Comment 3289: A strong case against "Ty Coon"

It would make sense that Richard Stallman may have chosen a word like "Ty Coon," as it would be hangover from the common vernacular used by the post WW1 generation. However, the word is used less often now, and is generally applied to "old money" professions, such as "oil tycoons." The word it would seem was originally meant tongue-in-cheek as it is now. It was derived from the Japanese word 大立者, and used as a way to impress US invaders of Japan in the mid-19th-century. However, this does not seem like a lasting word, and certainly not one that a person can derive from greek or latin origins, thus, it will only be all the more confusing for non-native speakers to derive the meaning. Lastly, by splitting it up into "Ty Coon," it may even have a worse interpretation, namely that someone not understanding the word 'tycoon', with its tricky etymology, may look up "Ty" and "Coon." Ty being an uncommon word, (language code for tahitian, shortwave for "they") they would be left with Coon. Perhaps they would think that "Coon" meant the animal or perhaps they would look at its just as common meaning "disparaging term for a Black person."

Comment 3290: Yet another reason to remove the date

Quote from

http://www.itwire.com.au/content/view/12685/1090/

"Microsoft has been emboldened by the fact that the last call draft of GPLv3 does not penalise Novell. A bit of compromise on Richard Stallman's part has been interpreted as weakness."

Lets put it this way: had the situation been reversed, would Microsoft show any mercy at all? *NO!*

They are a greedy corporation, a corporation that is comfortable suing members of its own development community (http://www.theregister.co.uk/2007/06/05/microsoft_mvp_threats/), comfortable stifling competitors and innovation, so I doubt they will hesitate exploiting this new soft underbelly in GPL3.

The argument about coupon distribution aside, this date gives them much more room to wriggle and squirm out of GPL3 obligations.

The FSF has been, and still is, the pillar that safe-guards free software developers' rights to their fullest, and it should always continue to do so. By allowing possible loopholes to be exploited, that pillar just doesn't seem so strong anymore...

If you read their rationale, the FSF's position is that they can use the Microsoft-Novell deal to the community's advantage and Microsoft's detriment. I still think it's a mistake to allow Novell to remain part of our community by giving them what amounts to an explicit exception to the rules of the GPLv3.

Exactly, skquinn. Novell went behind the back of the community and deliberately exploited the wording in the GPL2 to be enter the patent-conevnant with Microsoft.

They should not be rewarded for what they did! Which is what this date is going to do if its not removed... There should be absolutely no exceptions. Letting Novell continue with their deal simply makes it look like, from the corporate perspective, that they are the only "legitimate" GNU/Linux distributor to be used in a business environment.

Excuse the typos:

"to be enter" -> to enter "patent-conevnant " -> patent-covenant

I don't think that the GPLv3 should be "aimed" at any particular company or organisation. I am not concerned with Microsoft, or Novell.

I certainly wouldn't want to "show no mercy" just because Microsoft would "show no mercy". After all, aren't we better than Microsoft?

I have my own views on what this date should be, but that is in another comment.

(As an aside, concerning Novell, take a look at the June 2 2007 news on the following link. Look at the list of contributors. Should this be acceptable? Would this be permitted under GPLv3? Or should Novell be allowed to contribute, but not distribute?

Or maybe life is never clear-cut!

http://gcc.gnu.org/ )

jothal

Comment 3291: Overreaching?

One question before GPLv3 becomes final: Are you sure this section is not overreaching? For example:

A Creative Commons license could be considered a measure to protect the rights of an author (no commercial use, say) while still allowing sharing. Using a covered work to securely record this license and make it available when the work is accessed could be construed as a technological measure to secure these rights under WIPO article 11. Would this section not prohibit the use of GPLv3 covered works in this beneficial activity?

Comment 3293: For *humans*

What if a project had some binary blob-ish files that they preferred to maintain by using a program to process those blobs whenever they needed changes.

Maybe this could be avoided by mentioning humans somewhere in the definition? "the form preferred for humans to understand and make modifications to it"?

This fixes the situation where there are three forms used in the development process: a) A text form b) A post-processing binary-ish form c) The final binary output

Where the project prefers to perform maintenance on the data in form (b) rather than (a), we have to make sure that the definition still requires that the distributor makes (a) available, not (b). So adding something about human comprehension seems to protect against this.

Comment 3294: By name or just the fact?

If I modify a work, do I have to give my name when saying that I modified it?

Can I say: "I modified this on $date"

Or do I have to say: "Modified by Ciaran O'Riordan on $date" ?

So, can changes be made anonymously and distributed, or not?

Perhaps this should be written as: "a) The work must carry prominent notices identifying the owner of the copyright on the modifications and giving a relevant date."

I agree, "identifying the owner of the copyright..." would be better than just "you".

Comment 3295: and "making available to the public"

A European lawyer has told me that this should say:

"The precise terms and conditions for copying, distribution, making available to the public and modification follow."

This would internationalise the text in the same way and for the same reasons as the definition of the term "to propagate", and is necessary, I'm told, because recent changes in European laws have modified because the concept of "distribution".

Comment 3296: But what does that mean?

At what point does code fall under the definition of "entire work" when it is someone else's software?

If "libsomething" were a GPL3 library, and my program that is proprietary linked to it, what makes that a violation? At what point is the actual "entire work" both my application and libsomething?

For that matter, the GPL2 had the same problem as it didn't define "derivative work".

Comment 3297: But what does that mean?

At what point does code fall under the definition of "entire work" when it is someone else's software?

If "libsomething" were a GPL3 library, and my program that is proprietary linked to it, what makes that a violation? At what point is the actual "entire work" both my application and libsomething?

For that matter, the GPL2 had the same problem as it didn't define "derivative work".

Comment 3298: Define link

"link" is actually a technical term that should be defined somewhere in the license, preferably section 0.

Comment 3299: artificially splitting the distribution

if an "extension" (whatever that might be) is not covered work, but "interacts" with covered work, why is the developer forced to split up the useful combination into different distribution mediums just to circumvent the GPLv3 to be applied to his not-covered work? This just reduces usability for the users.

It's also confusing, and complicates things by creating a third class between "covered" and "not-covered", with potentially its own border-line cases. It's the extra confusion and complication I object to.

Comment 3300: There are legitimate "lock down" uses

There are legitimate reasons to "lock down" some hardware, e. G. calibrated measuring devices, like water/gas meters, or the IBM/Norwich Union Pay as you Drive devices.

The whole business case for both the conveyor and the user is based on the fact that the user does not change / manipulate the device software.

And as forbidding fraud does not prevent it, those devices will contain physical precautions to enhance tamper-proofness. Nevertheless, the conveyor may have the possibility to update the software, e. G. via cryptographically secured network access.

How do you cope with those cases?

These sounds like they might be situations where ownership of the device was never transferred to the user (water meters, for instance, typically belong to the water company, at least in my part of the world). If the device belongs to the company, it's not conveying the software, so those aspects of the license don't apply.

[Note that I personally oppose including anti-TiVoization in the license, because I believe specialized devices are not and should not be treated as general-purpose computers; however, it seems that there is a community that doesn't want its work used that way and I believe it should have the right to license its work as it likes...]

Comment 3301: files or code?

Neither "source" nor "source files" have been defined, so maybe this should be "source code" or "source code files"?

Comment 3302: Don't introduce a new word

Instead of "link", say "combine"; otherwise you limit yourself to software and to particular software technologies that include linkers. You also leave open claims to interpret "link" in the hypertext sense.

Comment 3303: could it be broader?

This is quite narrow. Are Blackberry mobile email checkers normally used for personal, family, or household purposes?

I couldn't argue they aren't. They are normally used by the execuatives of companies.

Our top-end routers? or high-end servers? Or desks with built in electronics? Or Field-programmable gate arrays? Or the hardware that hackers like the GNU Radio folk play with?

How about adding "or a use which can be reasonably anticipated" - to cover what hackers do. What hackers do can often be reasonably anticipated, but would not necessarily be a common or normal use.

Comment 3304: UNDEFINED patents

Why not put something in the GPL that prohibits cross-licensing patent agreement where the patents in the agreement are “UNDEFINED”. This seems to be Microsoft’s strategy and doing this will close one more loophole.

Comment 3305: Coverage is undefined

The term "coverage" has not been used previously and it's not at all clear what it means here. Instead, say, "doubtful cases shall be resolved in favor of considering the product a consumer product."

Comment 3306: Too broad?

Saying "a typical or common use" suggests that there may be more than one typical use of the device and implies that they can be considered separately, even though the end of the same sentence says they can't.

This whole paragraph reeks of insufficient review and rewriting - it jumbles a bunch of semi-independent thoughts together without really blending them into a coherent principle.

Comment 3307: Affero GPL Proxy

Maybe one should allow a proxy for the Affero GPL "or any later version" option.

This might not as important as only the work that originally was under Affero GPL is covered by the Affero GPL and I cannot think of anything bad that could happen to "your" work in consequence of this.

Comment 3308: Add an exeption

I think software license never dominates hardware.

In addition, in some cases, especially for product safety, the requirement of providing Installation Information may not be appropriate.

As a middle course, an exeption, described below, should be added.

"if the User Product has a notice, on the package and/or the web page, so as to the recipient can read it before buying, which shows (1) the User Product uses software licensed under the GNU GPL and (2) it is not designed for enabling to install and execute the modified version."

Such notices will enlighten general public on software freedoms. This will be a great campaign.

Comment 3309: Can AGPLv3'd library add additional requirements into GPLv3'd application?

If there is a non-interactive library under AGPLv3, and someone links/combines it with GPLv3'd web front-end, does the front-end need to provide a way to get the corresponding source of the AGPLv3'd library? If it does, the requirements of AGPLv3 introduce an additional restriction into GPLv3'd code. It's more than allowing plain "linking", so maybe AGPLv3'd libraries with GPLv3'd front-end would be problematic in legal point of view.

Instead of requiring the front-end to make the source code available, AGPLv3 could simply require that anyone using the AGPLv3'd work on a public server would be required to make the source available. That is, move the responsibility from the front-end author to the administrator of a public server. The only problem I see is that the administrator isn't required to accept GPLv3 nor AGPLv3, which could make this suggestion too weak.

Comment 3310: For clarification

The word "required" should be changed with "essential".

If the receipient does not have enough skill, the Installation Information goes far beyond the intention.

Comment 3311: comma needed after Fifth Floor

A comma is needed after the "Fifth Floor" otherwise the city could seem to be "Fifth Floor Boston". American readers will figure it out anyway, but for internationalizing the license, make it easy for everyone else too.

And a period at the end (just before the word "Everyone") is in order too.

Comment 3312: Free programs may not be GPLed

These pieces will be usable only in GPLed programs. According to FSF documents and general understanding, free programs are not necessarily GPLed programs. This statement can therefore be interpreted as totalitarian, and will certainly lead to confusion and conflicts. Ennemies of free software are always trying to confuse issues by distorting the meaning of words, or misusing words. We should neither help them, nor be guilty of similar behaviour.

Comment 3313: generalize beyond software

This license applies beyond software, and the selected phrase can be rewritten to reflect that virtue: "freedom to share and change all versions of a work--to ensure that it remains free for all recipients"

Comment 3314: generalize beyond software

To generalize beyond software, change programs => works

Comment 3315: slight edit

make sure => ensure

Comment 3316: clarity is in the eye of the beholder

Just say "the GPL explains". Whether it's clear is up to the reader!

or, IMHO even better, "states"

Comment 3317: This will weaken the GPL

As a programmer and promoter of free intangible creations, I am quite happy to be forced to waiwe any legal power I may have to prevent circumvention. None of my business. However, if local law requires the use of an effective technological measure, I will just have to switch to another licence in order to implement such measures in free software without legal risk to myself for unlawful circumvention. While the second paragraph seems OK, at least as I read it, the first paragraph can only hurt the GPL and GPLed code, for no concrete result.

Currently, if I want to read some CDs or DVDs on my machine, I am legally obliged to have such technical measures in the software. If the GPL licence forbids it, I will just have to use another licence that allows it. Normally GPL dominates the market. But when any GPLed code is asserted to be illegal by the GPL itself, any GPL implementation will be illegal and removed from the market, leaving all the needed space for other licences.

The problem is not the use of another license to release software that is part of an anti-circumvention mechanism, but the problem of such laws being introduced after software has been produced under the GPLv3.

The answer might be to "reimplement", but what (say) about larger projects where this might not be practical or even feasible?

The answer might be to use non-GPLv3 licenses in cases where there may be uncertainty, not because of what is the case now, but what might happen in the future.

jothal

Comment 3318: need a noun

"This" needs a noun after it for clarity. Perhaps "This asymmetry"? "This denial"? "This design"? The noun helps the reader link to the previous sentence or thought.

Comment 3319: unneeded qualification?

Is "to use" a necessary qualification? If not, just delete it (otherwise the reader wonders whether it is a necessary qualification).

Comment 3320: "states" is legally ambiguous

In most countries, "states" refers unambiguously to the national state or government; but in the US it could also mean states like the 50 states. Would "jurisdictions" be better here?

Comment 3321: "this" needs a noun

"this" needs a noun after it, to help the reader make the link. Perhaps "this outcome"

Comment 3322: copyright in

Isn't it "copyright in" (rather than "on")? In England it would be, but I'm not sure about the USA.

Comment 3323: is the intention important?

Is the intention important? Or is it bad no matter how it happens? In which case delete the "are designed to".

Comment 3324: just one work at a time

The licenses are for a single work at a time (where a collective work is one work), so I think "work" sounds better here than "works".

Comment 3325: In contrast

"In contrast" sounds better than "By contrast".

Comment 3326: generalize beyond software

Replacing the phrase with "When we say free" no longer limits the preamble to software.

Perhaps "free works".

Comment 3327: slightly unclear reference

I'd replace "them" with "the copies", for 100% clarity. I was wondering why it bothered me as "them" and I think it's the possibility of "them" referring back to Licenses, which is the other plural in the sentence.

Or perhaps "them" could be taken as referring to free software rather than copies of it.

Comment 3328: generalize

"software" => "work"

Comment 3329: jurisdictions

Is it always countries? Maybe use "jurisdictions" instead.

Comment 3330: need a blank line here

this subsection needs a blank line before it. Or is it supposed to be tightly merged with (b)?

Comment 3335: Could legally-granted moral rights interfere with and/or weaken the GPLv3?

The Creative Commons CcWiki has mentioned an issue with regard to legally-granted moral rights. For more information, please see http://wiki.creativecommons.org/Version_3#International_Harmonization_.E2.80.93_Moral_Rights (checked on June 10, 2007.) In particular, the page says that an author's moral right of integrity can be very restrictive in the jurisdiction of Japan. More specifically, Japan's moral right of integrity can supposedly be interpreted in a such a manner as to disallow _any_ alterations to a creative work. For Creative Commons, this posed a problem because some of the Creative Commons licenses allow users to produce derivative works. The solution was to limit the moral right of integrity for jurisdictions where it could completely disallow modifications to CC-licensed works. The unported (not specific to a certain jurisdiction) version of the Creative Commons Attribution-ShareAlike 3.0 license has a clause along these lines in its legal code. Please see section 4(d) at http://creativecommons.org/licenses/by-sa/3.0/legalcode for details.

On the same CcWiki page, it is mentioned that some jurisdictions consider moral rights to be of great importance. For the jurisdictions that fall in this category, it is said that a license may need to recognize moral rights in order to be enforceable.

In the GPLv3, it might be useful to recognize any moral rights that exist, and to limit any moral right of integrity in the event that it is too restrictive. This would be similar to the policy for the unported Creative Commons Attribution-ShareAlike 3.0 license.

It might be worthwhile for the FSF to look into the issue of moral rights for various different jurisdictions.

Comment 3336: Could the statement about the threat of software patents be seen as being an opinion of the FSF?

For some, the statement about the threat of software patents might be seen as being an opinion of the Free Software Foundation as opposed to a factual statement. To be sure, there might be documented cases where patents have interfered with software development. It might be useful to consider the wording of the statement and to reword the statement, if necessary, so that it is clear as to whether the statement is an opinion or not.

Yes, I'm sure there has been some program (if only a trivial one), somewhere, at some time, that was (at least for a short period) not threatened by software patents. This sentence is probably a bit over the top.

Comment 3337: The statement about states and software patents is presumably an opinion-should it be reworded so that this is obvious to the reader?

Consider the statement about states and software patents. Presumably, this statement is an opinion of the FSF. In this case, it might be useful to change the statement so that it is obvious to the reader that the statement is an opinion.

Yes, there are probably some contributors to GPL works who do not actually agree with this (e.g. corporations), but it seems here that the license is speaking on behalf of all contributors.

Comment 3338: Is it appropriate for a serious license to include humor i.e. "Ty Coon"?

Given that the GPLv3 performs a serious function in terms of users' freedoms, is it appropriate for the license to include humor?

Comment 3340: Internal conflict?

The first paragraph has "Copyright (C) 2007 Free Software Foundation" and "but changing it is not allowed". The preamble states that the license is free and copyleft?

Yes, the GPL is actually a non-free copyright license, that is designed to make other works free and copyleft.

Comment 3341: Disallow changing

I don't think people should be allowed to modify the GPL as they may do so conveniently in order to take away other's freedoms or win a case in court. Maybe the GPL can stand as a publicly-created but thereafter immutable license.

By "modify" it means create a derived work. Any derived work is obviously not the GPL itself, so allowing the right to "modify" would not actually put the GPL in any danger of being modified as such. (Don't blame me, it's the FSF's terminology.)

Comment 3342: Warranty

The major problem I see with this concept is the enforcement of a warranty. For example a manufacture makes a Router which uses GPLv3 Software. According to this clause of the License, the said manufacture would have to allow all its users to install modified version of the firmware on the device. Which can "brick" the device, and for users to call the company for Warranty repairs. Costing said company huge amounts of money, also because their firmware update is most likely outside of Warranty coverage, still the Manufacture has no way to proof that given update happened, and caused the failure. It would be much better for users and for manufacture of the device to check the Digital Signature of the update, and only accept authorized updates. This is currently prohibited by this license. However a user of said device can call the manufacture, and ask them for the key or some other way to bypass the security check. The company would give user that key, but also void his warranty. Also there could be a toggle switch on the board to disable the Signature check, opening up the device would also void the warranty. This set-up would be a win/win for all players involved: User can hack the device all he wants as long as the company doesn't have to honor its warranty if something due to firmware hacking gets broken. However current license doesn't seem to address this concern for company’s legitimate need to stop people from flashing firmware on the devices, and then demanding warranty repairs if there is a failure. Thus, it would make sense to modify this clause to permit the Digital Signature check, as long as there is a known backdoor made by the manufacture to bypass it .

The next-to-last paragraph of section 7 gives the manufacturer the right to terminate the warranty service if the user modifies the device.

Yes. However, a firmware update would not leave any evidence that it came from manufacture or was installed by user. Thus, a user can always claim that "it was official firmware update that bricked the device" and demand manufacture to honor the warranty.

If the device would "force" the user to open it before he/she can flash it with "unauthprized" code, the manufacture can always show that device was physicaly tempered with.

Comment 3344: Patents

The case of a manufacture producing a security device, which runs software which implements a patented process. The let say the company would like to publish this software under GPL for reasons of security. This way people can inspect the firmware for any vulnerabilities. However, the device manufacture would not want to have to license, also the patent, to everyone who gets their hands on the GPL'ed firmware. This would be currently prohibited by this clause. This is unfortunate because, the device manufacture doesn't do anything against the software being Free, anybody can read/modify and run the software they simply can't infringe on the patent. The buyer of the device,would not infrige on the patent since he is a licensor. Nor would security resercher who runs the software in his lab. Only mass production would infringe on patent. Thus this license clause leaves company unprotected ,either forcing them to use another Open Source license or keeping the code closed.

Not licensing the patents to everyone is doing against the software being Free, because they cannot run the program without patent license. GPLv2 already prohibits the situation you are describing, and GPLv3 won't change it at all.

Comment 3345: Is the software an ETM?

From reading that WIPO treaty, it's not clear that decoder-software would be "part of an effective technological measure" (ETM) in every implementation.

Maybe it is in the DMCA and in some EU member states, but maybe more general wording could make this more future-proof and international.

For example, what if it also said that no covered work could be declared "authorised" without that declaration being irrevokable and unconditional (including allowing modification without losing the "authorised" declaration)?

Comment 3346: Does this allow email?

Is it acceptable to provide a copy of the source by email? This paragraph does not appear to allow it in cases where you do not collect your mail from a server but receive it directly.

Comment 3348: Benefiting from and helping to realize third party patent claims

Does this section address a case where, for example, Microsoft gets access to Novell code destined to go into a GPL3 code base? Microsoft can study the code to come up with patents improving its chances of having no prior art and that code in FLOSS will exist that is infringing (because of Novell contributions made after MS files the patents). I think this case is a variation of a scenario I wrote up in a comment to "Section 11" on the GPL3draft3 comment page. I don't have the skill or time to analyze the GPL3 draft but wanted to help with potential loose ends.

Comment 3350: The definition of a "Major Component" is somewhat ambiguous

It appears that there at least two possible interpretations for this sentence. The first interpretation is that a "Major Component" is _one_ of the following: (a) A major essential component of the specific operating system for the executable work, (b) A major essential component of a compiler that produced the work, or (c) A major essential component of an object code interpreter that runs the work. The second interpretation is that a "Major Component" is _one_ of the following: (a) A major essential component of the specific operating system for the executable work, (b) One or more entire compilers that produced the work (c) One or more entire object code interpreters that run the work. Which of these interpretations, if any, is correct? It would be useful to reword this sentence so that it is more clear.

Comment 3351: What exactly is allowed with regard to "System Libraries" and "Major Components"?

Consider the situation where a GPL-licensed program is compiled with a proprietary compiler for use on a proprietary operating system. Presumably, it would be legitimate for the GPL-licensed program to use the standard APIs for the operating system (i.e. for creating windows.) If a proprietary programming environment supplies non-GPLed code for the purpose of accessing the operating system APIs, would it be legitimate for a GPL-licensed program to statically incorporate that code? Would this code fall under the "System Libraries" definition? According to the GPL, a System Library must be "included in the normal form of packaging a Major Component..." among other things. Given this requirement, it make a difference if the operating system and the compiler were sold as separate products from separate manufacturers?

If a BASIC compiler uses proprietary runtime code to implement the underlying functionality of the various BASIC statements, can a GPL-licensed program be legitimately compiled with that compiler?

Comment 3352: Potential for abuse?

A cellular company seems to be able to upload modified GPL software to phones without giving users full GPL rights.

A cellular company might distribute customer care software based on GPL software to resellers without giving them full GPL rights.

Could tivo sell empty machines and download the software to them bypassing the GPL?

If the interpretation of these clauses will be open to debate in court, I would rather see them gone.

Under GPLv2, the distributing party was prohibited to impose additional restrictions (ie: you have to accept this confidentiality agreement), but now it seems GPLv3 offers a mechanism to do just that: "under your direction and control, on terms that prohibit them from making any copies of your copyrighted material outside their relationship with you".

I have a non-academical very real world interest in this matter regarding software for which I hold copyright, that will regrettably automatically upgrade to v3. If you think this comment merits a response, please contact me through my registered email. Thank you very much.

Comment 3353: Is ROM a good example of unmodifiable code?

Consider a device that has GPL-licensed code in the form of a ROM chip. Even though the contents of the chip cannot be modified, a user might be able to substitute a different ROM chip or an equivalent medium that mimics the ROM chip. In this way, a user might be able to have the device use different code. Soldering the existing ROM chip (as opposed to socketing the chip) would not necessarily prevent this. If it is theoretically possible for the user to substitute different code in this manner, should the requirement to provide Installation Information apply to ROM-based code?

Comment 3354: Could a future technological advance enable the modification of code that was previously "unmodifiable"?

Consider a situation where the Installation Information is not provided because it is figured that modified code cannot be substituted. This would appear to be legitimate. If a future technological advance enables users to substitute modified code on a device that predated the advance, can the manufacturer be liable for not originally providing Installation Information with the device?

Comment 3355: For free software

The GPL can't prohibit this practice for all software. It can prohibit this practice to be carried out with free software. This needs some clarification, the reader might be lead to think that the GPL will prohibit this practice in all cases. A suggestion: "... prohibit this practice for those products if those products use free software.

Comment 3356: Hey all,

online comments didn't work (ubuntu gnu on tpad).

from the Preamble: The licenses for most software and other practical works are designed to take away your freedom to share and change the works. By contrast, the GNU General Public License is intended to guarantee your freedom to share and change all versions of a program

The pagragraph starts with "works" and switches to "programs" which is confusing.

Perhaps the paragraph should start: The licenses for most software *programs* and other practical... --adrian custer

Comment 3357: The first sentence is an improvement from v2: it gives a good overview.

However, the subject being covered needs to be better defined. The rest of the preamble interchanges "software" "programs" and "works".

Perhaps "software" should be "software programs"?

Also "kind of works" gets echoed immediately by "practical works" in the next paragraph. Do we want the first reference to flesh out a more generic category which includes 'software" for example "practical works of authorship" or some similar construct? If the words are all used together early to show they are the same, then the rest of the preamble can use them interchangeably without confusion.

--adrian cuater

Comment 3358: In the Preamble,

that you can change the software or use pieces of it in new free programs

I would change this to read: that you can change the software or reuse parts of the software in your own programs

"in new free programs" should be dropped. Technically, the freedom to 'mashup' exists privately for any kind of software. It only needs to be *free* if it is re-distributed which is what the next paragraph is about.

--adrian custer

Comment 3359: >From the Preamble:

if you distribute copies of the software, or if you modify it

should read if you distribute copies or modifications of the software

The ", or if you modify it" is incorrect since anyone can modify the code for their own use without incurring any responsibility towards others.

--adrian custer

Comment 3360: Preamble, paragraph on embedded devices:

The systematic pattern of such abuse occurs in the area of products for individuals to use. Therefore,

this should be dropped or re-written if kept. This sentence is true but not necessary and therefore excessive.

The paragraph should read: Some hardware devices are designed to deny users access to install or run modified versions of the software inside them, although the manufacturer can do so. Because this arbitrary restriction is fundamentally incompatible with the aim of protecting users' freedom to change the software, we have designed this version of the GPL to

which makes the point but saves space. "arbitrary restriction" is in response to sanjoy's comment online that "this" is insufficient.

--adrian

Comment 3361: Preamble:

To prevent this, the GPL

whereas in the previous paragraph we had: we have designed this version of the GPL

so perhaps we should stick to using "this version of the GPL"?

--adrian

Comment 3362: Spread the word about freedom

This section should also provide a way to spread the word of freedom, e.g. by a file called FREEDOM written by Richard.

The addition could be:

"If you provide bundled software, and any of the programs contain the file FREEDOM in the source archive, you are required to give the user the option to view the content of the latest version of FREEDOM during installation for the bundled software."

Comment 3363: Consider removing the preceding "0. " from the "0. HOW TO APPLY THESE TERMS TO YOUR NEW PROGRAMS" title

It might be useful to remove the preceding "0. " from the title of this section. The previous section is titled "TERMS AND CONDITIONS" and there is no preceding number in that title.

Comment 3364: Allow providing source at market rate rather than cost

As far as I am aware, in reality, distributors of Free Software almost always convey source at reasonable market rate rather than reasonable cost. I think it would be a travesty if anyone were ever sued for distributing source at reasonable market rate, so this clause needs to be modified to make it fair, and to make it reflect reality.

(Note that by Free Software I mean software that is free in its entirety, like Debian, not software that includes free and non-free parts. Ironicly, I don't think this is a problem for the latter.)

I think people support "cost" due to a confusion between cost and market rate, which I address in the following FAQ-style dialogue:

/ Businesses will be happy with this, because they can recover their costs. /

When investors put money into a business, to buy land, build stores, buy equipment, train staff, etc., they expect that investment to make them a reasonable profit. Therefore, businesses only ever do things that they think will make a reasonable profit (in the case of making donations, they hope that doing this will improve their company image). Anything a business does which merely breaks even, is tying up capital that could otherwise be making a profit, and is reducing return on investment. Anyone who didn't want to make a profit would have used their money to start a non-profit organisation, or stuff a pillow. Businesses will not be happy.

/ Individuals are okay though, because they can charge for their labour. /

Sole-traders charge for their labour at market rate, not at cost. In fact, sole-traders don't record their cost of labour anywhere--they do not officially get paid wages, rather any money they make is considered profit. I'm not sure whether this means they couldn't charge anything at all for their labour, but they certainly couldn't charge a fair market rate.

/ "At cost" really means the "cost" of not getting the profit that you would have got by doing stuff at market rate, had you not been too busy distributing source at cost. /

If this is what it is supposed to mean, then it would be clearer to say market rate.

/ You could get around this by paying someone else to convey the source, and recording market rate as your cost. /

In theory you might be able to get away with this if you weren't able to convey, but if the "at cost" provision applies to you, then it is because you have conveyed a binary, so clearly you are able to convey. This means that even if you are paying someone else at market rate, you are still only allowed to recover cost (not market rate), so you have to fork out the rest from you own pocket.

The GPL does have a "solution" for this though... If you have a competitor who also sells the GPL work, then you can go to them and demand that they provide the source at cost, since the GPL requires them to. They could then bring the order back to you... (kind of like passing a hot potato)

/ You have already paid for the source when you pay for the object code. /

You don't actually pay for the object code, you pay to be distributed the object code. And why should you have to pay to be distributed the source code, if you don't actually want it? This means you are forced to subsidise other people.

/ GPL works are supposed to be free. /

This is supposed to mean free as in free-market, not free as in free-ride. If people are free to obtain the source at market rate for this service (or an equivalent service), then everyone is free. But if people can demand that others give them the work for 'free' (at cost), then no-one is free, because we are all obligated to provide free-rides. I know it is a cliche to say that the GPL is communist, but in the case of this clause, it actually really is.

Comment 3365: Scope for Private Copying Allows GPL Evasion

Last week, I finally realised that GPL V3 drafts have broadened the scope of propagation without conveyance to include the possibility that such propagation could exempt whole industries and whole professions of users from GPL protection. This could easily happen under the guise of internal propagation within an organisation because some trade and membership associations have such a great breadth of members that internal propagation in such organisations would be the equivalent of conveyance in all but name.

Draft 3 dropped the words "copy and" from conditions which had been required "to copy and convey" in sections 4,5,and 6. Dropping those words and related changes in the language for basic permissions in section 2 removed the possibility that such copying conditions might be required when propagating without conveyance.

There is a fairly simple remedy to the excessive scope for private copying in the GPL V3 draft which achieves the basic benefit of private copying without the risk of exempting large classes of users from GPL protection. Unfortunately, the remedy I have may seem to be inelegant because explicitly preserving appropriate space for private copying while avoiding the potential for abuse requires additional language specifying an exception to private copying. The issue of excessive private copying scope has a resolution but not by merely one or two simple elegant words. A little too much verbosity seems to be required to specify the exception. It would be unfortunate if language inelegance was an obstacle to protecting the GPL rights of the members of large organisations.

1. CHANGED LANGUAGE IN DRAFTS.

GPL 1 and 2 controlled copying to enforce user rights under the GPL and the space for private modification and copying without conveyance was merely an understood right left outside the explicit terms of the GPL. GPL V3 drafts have specified explicit terms for private modification and copying without conveyance.

Draft 3 changed basic permissions for propagation without conveyance in section 2 from "without limitation" to "without conditions". The one qualification in the draft to "without conditions" is "so long as your license otherwise remains in force." "There may be no significant semantic difference intended by the change from "without limitation" to "without conditions" except for less possible ambiguity in the word choice. However, the change removed the possibility of interpreting 'limitation' as if limitation imposed no restriction on the number of copies but might have conditions imposed by other terms of the license. The qualification of "so long as your license otherwise remains in force" could have no bearing on copying when "copy and" has been removed elsewhere in the license.

Dropping "copy and" changed the beginning of sections 4 for verbatim copying and conveyance, 5 for modified source copying and conveyance, and 6 for copying and conveying non-source forms.

Section 4 in draft 2 had started as follows.

"You may copy and convey verbatim copies of the Program’s source code as you receive it, in any medium, provided that ..."

Section 5 in draft 2 had started as follows.

"You may copy and convey a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4 above, provided that you also meet all of these conditions:"

Section 6 in draft 2 had started as follows.

"You may copy and convey a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways:"

At first glance, the changes which GPL V3 drafts introduce for the scope of private modification and copying seem to only explicitly specify what had been understood for the scope of private modification and copying without an explicit provision in GPL 1 and 2. However, if the effect of the GPL V3 draft changes are examined more deeply a serious hidden problem is revealed.

2. EXCESSIVE SCOPE PROBLEM.

Licensees may be either individuals or organisations which is just as it should be. One would ordinarily want organisations to be allowed to make private modifications and private copies without needing to observe conditions required for conveyance. Yet, applying the right to ignore conveyance conditions for private copying to the membership scope of some very large organisations stretches the meaning of private copying well beyond the intent of private copying.

When an organisation is a licensee, the organisation management may make private modifications and private copies of a program for the members of that organisation without having the propagation qualify as conveyance. The organisation members may not receive copies of Corresponding Source or even an offer for Corresponding Source. The members may not even receive any notice that a program is covered by the GPL or other terms specified by the GPL

When the members of an organisation are in accord with the management over the disposition of GPL software, there may be no problem. However, the pervasiveness of enormous organisations where management and members are not in accord over the disposition of GPL software is very problematic. I am not thinking here of large organisations where the members are employees of the organisation.

Freedom of private modification and private copying is certainly a freedom which must be preserved while guarding against organisations stretching the concept private modification outside of its intent.

2.1. ORGANISATIONS WHICH MAY STRETCH THE SCOPE OF PRIVATE MODIFICATION.

Such large organisations may be large trade associations and voluntary membership associations which often include every or most every member in a particular trade group or profession. Such organisations can constitute most of the market for some application domains. Even when these organisations are ostensibly governed by their membership, they often have a self perpetuating management group which functions in a top down manner often against the actual interests of the organisation members.

Businesses may treat their customers as members of an organisation. Each customer would be a member of a customer club, subscribers group, or some other customer organisation which the controlling business would manage for its members. Such customer groups might be preexisting organisations or they might be specifically created to exploit a loophole in the GPL if GPL V3 is promulgated using the current draft terms.

The full effect could be to make the impact of GPL software on some major markets negligible. Such large organisations would be able to treat GPL software as if it is proprietary to the full scope of their members. Once one such organisation was known to have violated the intent of the GPL while claiming safety in the actual terms of the license and surviving any challenge other similar organisations would take inspiration from that example. There would be other organisations seeking to exploit the loophole provided once it was seen as a safe harbour. Existing organisations would move to take advantage of the opportunity evade the GPL and businesses would move to form new customer organisations to evade the GPL.

You might hope that such organisations would not actually evade the spirit of the GPL in their practise. I may be hopeful enough to suppose that most such organisations would not violate the spirit of the GPL. However, it would take only one trade organisation or comprehensive membership organisation to exploit this loophole for the benefit of the GPL to be lost to an entire industry or profession of users.

I know specifically of some profession encompassing membership organisations which are basically proprietary software and publishing businesses disguised as beneficent organisations. These organisations have management practises constantly adverse to their members. These organisations observe the GPL only in the least way that they can manage, although, they would not violate the spirit of the GPL unless the terms of the GPL provided cover. Given that cover from the current draft, they would have little compunction about modifying GPL software and possibly combining it with their proprietary software. They would then copy that software for their members without informing their members that any GPL software was present. Members unaware that they might have had rights as GPL users could never attempt to exercise the rights of which they would be unaware. I will not publicly disclose the names of particular organisations about which I have knowledge because it would be impolitic for me to accuse organisations of actions that they have yet to take with GPL draft terms which I hope are not promulgated in their current form as part of an actual release.

2.2. WOULD SOCIAL PERSUASION CORRECT THE PROBLEM?

Social persuasion may be used to correct many problems which cannot be corrected within the license itself. However, the license must be strong enough to not encourage the violation in the first place. If the license is seen to provide terms endorsing an abuse, then the effect of an attempt at social persuasion will be substantially muted.

One might think that the actual people in the management of such organisations would evolve into taking a more cooperative approach. However, the evidence shows that management ensures that like thinking people are promoted to succeed other members of management.

Some membership organisations of which I am thinking particular have already been little moved by substantial vigorous complaints from their own members about their ongoing bad behaviour in other areas. These organisations have only grown stronger with regular bad behaviour, thus reinforcing their bad behaviour.

The license must be strong be strong first. The organisations of which I am thinking have a model adverse to the spirit of the GPL but would not violate explicit terms of the GPL.

2.3. WOULD ORGANISATION MEMBERS USE GPL SOFTWARE INDEPENDENTLY OF THEIR MEMBERSHIP?

Organisation members would be free to act independently of their organisation but would they do so in fact? During approximately the past 25 years, the restrictions of the proprietary software model have been dominant in the market and in most people's recognised experience of using software. Consequently, a very small portion of software users understand the personal and social good of using free software.

2.3.1. HOW MIGHT ORGANISATION MEMBERS CHOOSE GPL SOFTWARE INDEPENDENTLY?

Free software has succeeded in attracting large numbers of users mostly by appealing to the terms under which users evaluate proprietary software. Users evaluate proprietary software in terms of marketing strength, functionality, and cost.

Free software in general has grown terrifically in market strength and has come to dominate some markets, however, free software struggles to have any impact in many markets.

Free software often fairs favourably in terms of features as compared to proprietary software. Features of free software may not be undeniably better than comparable features of proprietary software. but free software often has important features which are not available in proprietary software.

Free software might have significant advantages over proprietary software in terms of no licensing cost. However, costs of training and maintenance are still real costs.

Some membership organisations ensure that the membership will support the organisation management and continue to pay membership fees by holding members to ransom for access to the members' own data and access to other data. Other organisations management often uses other coercive means to keep their members under control. Such practises already complicate an effort to bring free software to the members of these organisations without an additional adverse effect from a loophole in the GPL license.

Under the current draft terms, these very same organisations with coercive management would be able to get away with treating modified GPL software as if it is the management's own proprietary software. In terms that most users understand how to value, there would then be little differentiation between modified GPL software being presented to members as proprietary and GPL software offered independently. Worse, if private modifications combine the management's own proprietary software with privately modified GPL software all the features of both the proprietary and the GPL software could be presented as proprietary to the members. In terms that most users understand how to value, the differentiation between modified GPL software combined with proprietary software and presented to members as one proprietary program and GPL software offered independently would be adverse to GPL software. If the GPL software offered independently has no advantages which organisation members recognise initially, members will not choose independently offered GPL software. Members will never learn to appreciate the other advantages of GPL software if they are not first choosing GPL software in terms that they recognise.

2.4. WILL AGPL V3 SOLVE THE ISSUE?

Are the most problematic organisations not primarily using network services so that AGPL V3 can address the problem for covered software?

Certainly, the problematic organisations I know of primarily use network services software. Unlike the GPL V3 draft, the network use terms of the AGPL V3 draft would primarily be effective for the members of large membership organisations. (There are some problems with the AGPL V3 draft but those are the subject of AGPL specific comments.)

However, the worst of the problematic organisations which I know of also provide software to every member. The software copied for members is used in conjunction with the organisation's network service. The hazard from the worst of the problem organisations is that they would switch from providing software via network services to providing all software as part of a private modification, privately copied for all of their members; where the network use became only a source of data used by the private modification. A network service affected by AGPL V3 could then be transformed into what would be effectively proprietary software under the guise of private modification and private copying. The GPL V3 terms which are identical to AGPL V3 must not provide cover for evading AGPL V3.

2.5. IS THIS REALY A SOCIAL PROBLEM AND NOT A LICENSE PROBLEM?

Is the problem merely a social problem of bad management practises in large organisations and not something that a license should try to address? (The issue of whether organisations would be shamed into complying with the spirit of the GPL has been addressed above.) Such a question is similar to asking whether we should try to reform society, and the laws, including copyright and patent law, without having the GPL. We must of course attempt to reform society and the laws. The GPL is needed as a wedge to help promote social reform it is not a substitute for social reform.

People want to do what is good. People take choices contrary to the good primarily out of ignorance of which choices are good. The GPL is needed to point people in the right direction and to make that direction a requirement.

2.6. SEVERITY OF THE HAZARD.

Of course, no one can predict the outcome of terms which have never been tested in the world.

I cannot emphasise enough how important this issue is when the scope of some organisations membership is large enough to encompass entire industries and professions of users. I have been encouraging other developers to take an active interest in the GPL V3 drafting process from the release of the first discussion draft and have been speaking highly of the new terms that I have seen providing better protection for free software. This is the first issue that has given me any reservations about the benefits of the GPL V3 draft. I have seen no other draft changes as part of the GPL V3 drafting process which threaten to reverse the gains which GPL software has made in some markets.

3. SUGGESTED SOLUTIONS.

We want to preserve the benefit for private copying intended by the draft 3 removal of "copy and" from sections 4. 5, and 6 while avoiding the undue consequences in the current draft.

Terms for copying need to be restored with a special restriction on their effect. The words "to any party" may need to be added for convey to ensure that the qualification of organisation members only affects copying and not conveying.

Private copying is a basic permission granted by section 2 and would be unaffecteded except for the special qualification suggested for sections 4, 5, and 6.

3.1. EXCLUDE ORGANISATION MEMBERS FROM PRIVATE SCOPE.

One solution is to restore copying to the terms of sections 4, 5, and 6 for organisation members only.

This change may have the effect of including organisation employees under the terms of sections 4, 5, and 6. I personally favour disallowing organisation managers from taking away the rights of their employees under the GPL at work. However, organisation employers are not the issue about which I have expressed my concern. I would not want organisations management desire to disallow their employees' GPL rights for organisation private copies to be an obstacle to correcting the important issue. I have another suggestion to avoid the conflict over employees.

3.1.1. DIFF OF SUGGESTED CHANGE.

[[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}} Text from a section excluded for brevity of presentation is marked by unquoted ellipses, '...'.

"2. Basic Permissions."

...

" You may make[[,]] {{and}} run [[and propagate]] covered works that you do not convey, without conditions so long as your license otherwise remains in force. {{You may propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force; except that propagation to members of your orgazation is permitted soley under the conditions stated below for copying.}}" ...

...

"4. {{Copying or}} Conveying Verbatim Copies."

" You may {{copy to any member of your organization or}} convey {{to any party}} verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program."

...

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to members of your orgazation must follow the conditions of this section.}}"

"5. {{Copying or}} Conveying Modified Source Versions."

" You may {{copy to any member of your organization or}} convey {{to any party}} a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4 above, provided that you also meet all of these conditions:"

...

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to members of your orgazation must follow the conditions of this section.}}"

"6. {{Copying or}} Conveying Non-Source Forms."

" You may {{copy to any member of your organization or}} convey {{to any party}} a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways:"

..

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to members of your orgazation must follow the conditions of this section.}}"

3.2. EXCLUDE ORGANISATION MEMBERS WHO ARE NOT EMPLOYEES FROM PRIVATE SCOPE.

Another solution is to apply private copying conditions to the terms of sections 4, 5, and 6 for organisation non-employee members only. This avoids the conflict over employees in addressing the important issue.

3.2.1. DIFF OF SUGGESTED CHANGE.

[[Double brackets mark material to be deleted.]] {{ Double braces mark material to be added.}} Text from a section excluded for brevity of presentation is marked by unquoted ellipses, '...'.

"2. Basic Permissions."

...

" You may make[[,]] {{and}} run [[and propagate]] covered works that you do not convey, without conditions so long as your license otherwise remains in force. {{You may propagate covered works that you do not convey, without conditions so long as your license otherwise remains in force; except that propagation to non-employee members of your orgazation is permitted soley under the conditions stated below for copying.}}" ...

...

"4. {{Copying or}} Conveying Verbatim Copies."

" You may {{copy to any non-employee member of your organization or}} convey {{to any party}} verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program."

...

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to non-employee members of your orgazation must follow the conditions of this section.}}"

"5. {{Copying or}} Conveying Modified Source Versions."

" You may {{copy to any non-employee member of your organization or}} convey {{to any party}} a work based on the Program, or the modifications to produce it from the Program, in the form of source code under the terms of section 4 above, provided that you also meet all of these conditions:"

..

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to non-employee members of your orgazation must follow the conditions of this section.}}"

"6. {{Copying or}} Conveying Non-Source Forms."

" You may {{copy to any non-employee member of your organization or}} convey {{to any party}} a covered work in object code form under the terms of sections 4 and 5, provided that you also convey the machine-readable Corresponding Source under the terms of this License, in one of these ways:"

...

"{{You may copy covered works that you do not convey, without following the conditions of this section except that copying to non-employee members of your orgazation must follow the conditions of this section.}}"