Here's an update on some of what's been going on in the copyleft-next draft since a week ago.
One of the more significant changes is a substantial simplification of the termination section. For what are best summarized as "historical reasons", GPLv3 has effectively two cure provisions (interestingly, MPL 2.0's drafters chose to adapt much of the GPLv3 language). I've replaced it with one cure period and I just refer to a "reasonable time period" (cf. EPL 1.0) instead of the 30-day and 60-day periods of GPLv3. I'm not convinced that it's better not to have a fixed time period for cure, though I am convinced that a 30-day and a 60-day provision are not both needed.
The structure of having automatic termination coupled with automatic reinstatement upon cure is preserved. This is unusual but it reflected concerns expressed by those active in GPL enforcement in Germany that automatic termination facilitated enforcement under German law.
I have added a 'severability clause' which is a common feature of contracts and is found in a number of FLOSS licenses. The FSF opted not to have a severability clause in GPLv3 out of concern that it could lead to unpredictable judge-made alterations of the GPL that might frustrate basic policy objectives. The advantages of a severability clause seem to me to outweigh this concern. (GPLv2 has a limited severability clause in its 'liberty-or-death' section, which was removed from GPLv3.)
I added a section at the beginning giving a 'general rule of construction', in some ways replacing one function the GPL Preamble ought to serve. The language here was adapted from the Preamble of Allison Randal's experimental short GPLv3 alternative which she drafted in 2007.
I deleted the lengthy yet narrow provision I call, for lack of a better name, "child of downstream shielding" (GPLv3 section 11 paragraph 5). This is actually a very important provision for historical reasons, as a guide to understanding the FSF's interpretation of the GPL, but I believe it has proven to have little value in the license itself. With this deletion, note that all the patent-related provisions of GPLv3 are now gone except for the Licensor's patent license grant (which has been restructured but, at least as I conceive it currently, is not different in scope from the patent license grant of GPLv3).
Sadly, vanilla copyleft-next is still over 100 lines longer than the Apache License 2.0, but great progress has been made. :-)
- Richard
On 07/25/2012 11:09 PM, Richard Fontana wrote:
With this deletion, note that all the patent-related provisions of GPLv3 are now gone except for the Licensor's patent license grant
Sorry, actually that's not strictly true - the clause that indicates that initiation of patent litigation that accuses any part of what I currently call the Received Work (GPL 'Program') violates the condition against imposition of 'further restrictions', using language that deliberately tracks the patent termination clause of the Apache License 2.0, is still present. (The corresponding clause is at the end of GPLv3 section 10.)
- Richard
On Thu, Jul 26, 2012 at 6:45 AM, Richard Fontana fontana@sharpeleven.org wrote:
On 07/25/2012 11:09 PM, Richard Fontana wrote:
With this deletion, note that all the patent-related provisions of GPLv3 are now gone except for the Licensor's patent license grant
Sorry, actually that's not strictly true - the clause that indicates that initiation of patent litigation that accuses any part of what I currently call the Received Work (GPL 'Program') violates the condition against imposition of 'further restrictions', using language that deliberately tracks the patent termination clause of the Apache License 2.0, is still present. (The corresponding clause is at the end of GPLv3 section 10.)
When I have time, I will argue that MPL's stronger termination clause is superior to Apache's. At the very least, copyleft.next should make more explicit what occurs when further restrictions are imposed, and when the imposition of further restrictions can trigger the prohibitions. The current language can be read creatively to apply only to additional restrictions you impose on recipients you distribute to.
Or to put it another way: it should be much more clear what happens to your rights to use the code when you sue or threaten to sue someone using the same code who got it from a third party.
Luis
Luis
Richard Fontana wrote:
One of the more significant changes is a substantial simplification of the termination section. .... I've replaced it with one cure period and I just refer to a "reasonable time period" .. I'm not convinced that it's better not to have a fixed time period for cure, though I am convinced that a 30-day and a 60-day provision are not both needed.
And thus master branch of copyleft-next now reads:
However, if You cure such failure within a reasonable time period after becoming aware of such noncompliance, then Your rights are reinstated.
I strongly believe this will make copyleft-next de-facto unenforceable. We'd have to litigate it all the way to the bitter end to get a ruling that "reasonable time period" is less than the average life cycle of most electronics and software products.
In my experience, nearly all copyleft violators believe it's reasonable to come into compliance only after a product is off the market. And most would happily say "see you in Court" if a copyright holder says otherwise.
Even those violators who make vague, half-hearted attempts to come into compliance still believe heavily protracted schedules are reasonable. For example, I've got compliance actions open today that were first opened last fall, and I'm sure the lawyers would argue, that they're still "within a reasonable time period" and tell me to sue them if I disagreed (of course, I already do disagree). As everyone knows, because I know the community prefers that we avoid lawsuits, I desperately avoid them until there is just no other choice, but violators' lawyers now know this too, of course, and play chicken.
Since most of my compliance work is GPLv2, it means such violators are (of course) dead to rights. Think about it: even with GPLv2's termination clause, which everyone agrees very strict (from a policy perspective), most violators take nearly the entire product lifecycle to come into compliance, even when they know they've gotta beg for their rights back at the end of it or face continued copyright infrigement even when *in* compliance on all other non-termination points. If the termination clause says "reasonable time", what incentive will be left to get them into compliance, ever?
My point, simply: I think copyleft-next will have to chose between naming a specific number of days, or be effectively unenforcable. Also, realize that *whatever* time frame you pick will be the minimum time by which we get code from companies gaming the system with delayed releases (which is already common, even with GPLv2).
-- bkuhn
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 07/30/2012 07:35 PM, Bradley M. Kuhn wrote:
Richard Fontana wrote:
One of the more significant changes is a substantial simplification of the termination section. .... I've replaced it with one cure period and I just refer to a "reasonable time period" .. I'm not convinced that it's better not to have a fixed time period for cure, though I am convinced that a 30-day and a 60-day provision are not both needed.
And thus master branch of copyleft-next now reads:
However, if You cure such failure within a reasonable time period after becoming aware of such noncompliance, then Your rights are reinstated.
I strongly believe this will make copyleft-next de-facto unenforceable. We'd have to litigate it all the way to the bitter end to get a ruling that "reasonable time period" is less than the average life cycle of most electronics and software products.
In my experience, nearly all copyleft violators believe it's
reasonable to come
into compliance only after a product is off the market. And most would happily say "see you in Court" if a copyright holder says otherwise.
Even those violators who make vague, half-hearted attempts to come into compliance still believe heavily protracted schedules are reasonable. For example, I've got compliance actions open today that were first opened
last
fall, and I'm sure the lawyers would argue, that they're still "within a reasonable time period" and tell me to sue them if I disagreed (of course, I already do disagree). As everyone knows, because I know the community prefers that we avoid lawsuits, I desperately avoid them until there is just no other choice, but violators' lawyers now know this too, of course, and play chicken.
Since most of my compliance work is GPLv2, it means such violators are (of course) dead to rights. Think about it: even with GPLv2's termination clause, which everyone agrees very strict (from a policy perspective),
most
violators take nearly the entire product lifecycle to come into
compliance,
even when they know they've gotta beg for their rights back at the end
of it
or face continued copyright infrigement even when *in* compliance on all other non-termination points. If the termination clause says "reasonable time", what incentive will be left to get them into compliance, ever?
My point, simply: I think copyleft-next will have to chose between naming a specific number of days, or be effectively unenforcable. Also,
realize that
*whatever* time frame you pick will be the minimum time by which we
get code
from companies gaming the system with delayed releases (which is
already common,
even with GPLv2).
-- bkuhn _______________________________________________ copyleft-next mailing list copyleft-next@lists.fedorahosted.org https://lists.fedorahosted.org/mailman/listinfo/copyleft-next
At this point I am inclined to agree with Mr. Kuhn and favor a time period being included to ensure enforcement. A tight timetable is preferred especially in today's environment where stock markets can swing wildly day by day due to the release of any news. There must be an incentive for compliance whether it be a carrot or a stick.
Stephen Michael Kellat
On Mon, Jul 30, 2012 at 7:35 PM, Bradley M. Kuhn bkuhn@ebb.org wrote:
My point, simply: I think copyleft-next will have to chose between naming a specific number of days, or be effectively unenforcable.
I agree that this is the likely practical outcome. I generally favor a short period, but there is one corner case that concerns me. Imagine Company X has failed to comply with copyleft-next and loses its rights to incorporate Some Software Package in its product. After six months, the board of directors decided the legal and PR exposure of such a stance (yes, we're deep into the hypothetical woods at this point) is unacceptable and sacks the executives. The new executives immediately move to come into compliance with copyleft-next. However, by this point the reinstatement period has passed so Company X still cannot use Some Software Package. (Presumably, a new release of SSP would give Company X the chance to prove their commitment to copyleft, but maybe by that time SSP is no longer under development.)
To put it in more general terms: if the circumstances that cause an actor to be a bad actor change such that the actor is now a good actor, it may be too late. This strikes me as being punitive and not in line with what I imagine the ideals behind copyleft to be (this ties back in to the "platonic ideals" thread). While I definitely agree that people who violate the license should lose their rights under the license, I also think we should be willing to accept atonement, however delayed it might be. Now how to do that in a practical and legal manner such that it doesn't result in licensors getting walked all over...well, I haven't a clue.
On 07/30/2012 07:35 PM, Bradley M. Kuhn wrote:
Richard Fontana wrote:
One of the more significant changes is a substantial simplification of the termination section. .... I've replaced it with one cure period and I just refer to a "reasonable time period" .. I'm not convinced that it's better not to have a fixed time period for cure, though I am convinced that a 30-day and a 60-day provision are not both needed.
And thus master branch of copyleft-next now reads:
However, if You cure such failure within a reasonable time period after becoming aware of such noncompliance, then Your rights are reinstated.
I strongly believe this will make copyleft-next de-facto unenforceable. We'd have to litigate it all the way to the bitter end to get a ruling that "reasonable time period" is less than the average life cycle of most electronics and software products.
In my experience, nearly all copyleft violators believe it's reasonable to come into compliance only after a product is off the market. And most would happily say "see you in Court" if a copyright holder says otherwise.
Even those violators who make vague, half-hearted attempts to come into compliance still believe heavily protracted schedules are reasonable. For example, I've got compliance actions open today that were first opened last fall, and I'm sure the lawyers would argue, that they're still "within a reasonable time period" and tell me to sue them if I disagreed (of course, I already do disagree). As everyone knows, because I know the community prefers that we avoid lawsuits, I desperately avoid them until there is just no other choice, but violators' lawyers now know this too, of course, and play chicken.
[...]
My point, simply: I think copyleft-next will have to chose between >
naming a specific number of days, or be effectively unenforcable.
This is precisely the reason (well, one reason) why I hoped you might post to this list, since there are very few people in the world other than you who have dealt with the practical issues surrounding GPL enforcement (on the licensor/copyright holder side, of course).
Do you have views on what fixed period is 'default reasonable'? 60 days? As I've said, I don't want a 30-day and 60-day dual scheme, since I can't see how the added complexity is justified. Review of the history of GPLv3 drafting shows that the FSF started out with just the 60-day repose idea. I will check into any public archives relating to MPL 2.0 drafting to see why Mozilla decided to adapt the 30/60 day dual approach for MPL 2.0 (or if Luis is reading this he may be able to explain or point to something :).
- Richard
On Mon, Jul 30, 2012 at 7:08 PM, Richard Fontana fontana@sharpeleven.org wrote:
On 07/30/2012 07:35 PM, Bradley M. Kuhn wrote:
Richard Fontana wrote:
One of the more significant changes is a substantial simplification of the termination section. .... I've replaced it with one cure period and I just refer to a "reasonable time period" .. I'm not convinced that it's better not to have a fixed time period for cure, though I am convinced that a 30-day and a 60-day provision are not both needed.
And thus master branch of copyleft-next now reads:
However, if You cure such failure within a reasonable time period after becoming aware of such noncompliance, then Your rights are reinstated.
I strongly believe this will make copyleft-next de-facto unenforceable. We'd have to litigate it all the way to the bitter end to get a ruling that "reasonable time period" is less than the average life cycle of most electronics and software products.
In my experience, nearly all copyleft violators believe it's reasonable to come into compliance only after a product is off the market. And most would happily say "see you in Court" if a copyright holder says otherwise.
Even those violators who make vague, half-hearted attempts to come into compliance still believe heavily protracted schedules are reasonable. For example, I've got compliance actions open today that were first opened last fall, and I'm sure the lawyers would argue, that they're still "within a reasonable time period" and tell me to sue them if I disagreed (of course, I already do disagree). As everyone knows, because I know the community prefers that we avoid lawsuits, I desperately avoid them until there is just no other choice, but violators' lawyers now know this too, of course, and play chicken.
[...]
My point, simply: I think copyleft-next will have to chose between >
naming a specific number of days, or be effectively unenforcable.
This is precisely the reason (well, one reason) why I hoped you might post to this list, since there are very few people in the world other than you who have dealt with the practical issues surrounding GPL enforcement (on the licensor/copyright holder side, of course).
Do you have views on what fixed period is 'default reasonable'? 60 days? As I've said, I don't want a 30-day and 60-day dual scheme, since I can't see how the added complexity is justified. Review of the history of GPLv3 drafting shows that the FSF started out with just the 60-day repose idea. I will check into any public archives relating to MPL 2.0 drafting to see why Mozilla decided to adapt the 30/60 day dual approach for MPL 2.0 (or if Luis is reading this he may be able to explain or point to something :).
We saw no reason to deviate from what GPL v3 was doing in this regard; it seemed reasonable/workable and GPL compatibility was a significant goal for us at the time.
The two-stage approach is also reasonable, I think, because licensee-notice+cure is different from self-awareness+cure. I can understand why you'd want to slim it down to a single period, though.
Luis
On 07/30/2012 10:14 PM, Luis Villa wrote:
On Mon, Jul 30, 2012 at 7:08 PM, Richard Fontana fontana@sharpeleven.org wrote:
I will check into any public archives relating to MPL 2.0 drafting to see why Mozilla decided to adapt the 30/60 day dual approach for MPL 2.0 (or if Luis is reading this he may be able to explain or point to something :).
We saw no reason to deviate from what GPL v3 was doing in this regard; it seemed reasonable/workable and GPL compatibility was a significant goal for us at the time.
Ah, the principle that "mere nonsubstantive differences in termination provisions cannot lead to GPL incompatibility" never got codified (and may never have been fully thought out anyway).
The two-stage approach is also reasonable, I think, because licensee-notice+cure is different from self-awareness+cure.
Yes, very good point; I thought about this after posting that message. They really are two categorically different provisions. However that itself seems to be something that adds unnecessary complexity.
- Richard
Richard Fontana wrote at 22:08 (EDT) on Monday:
Do you have views on what fixed period is 'default reasonable'? 60 days? As I've said, I don't want a 30-day and 60-day dual scheme, since I can't see how the added complexity is justified.
IIRC, Dave Turner influenced that based on his enforcement experience working with me at FSF on GPL enforcement. Even though no GPLv3 drafters (except Fontana) ever bothered to ask my opinion on any part of GPLv3, I always rather liked GPLv3§8, notwithstanding the complexity.
Note that the "provisionally, then 60 days" thing ensures the copyright holder can enforce effectively against the perennial violator who self-corrects *only* when a user notices but falls out of compliance again quickly.
I also somewhat get why the second number must be 30 days (i.e., shorter), because the copyright holder has to ramp up legal action *after* the termination is permanent, so waiting 30 days really means waiting 90 or even 180 to get a Court to issue an injunction, since the 30 day mark is merely your *filing* date.
Anyway, as I rereview GPLv3§8, I'm again convinced all moving parts are necessary. I can take any given part and explain why it's good enforcement policy (I've not covered it all above, but ask me about specific clauses and I'll answer).
But, that all said, I've never actually enforced the GPLv3. All my enforcement has been GPLv2, and thus I've never enforced under any termination clause other than "auto-termination on any infraction and you'll just have to beg for your permissions back". I've often wondered if termination weren't automatic, violators would work harder to get back into compliance before termination happened. I wonder sometimes if they feel they're screwed anyway right now, so they might as well just sit around and wait for the lawsuit to come.
I've also wondered if different termination policies should exist for different types of violations. But that would just add even more complexity, so I presume copyleft-next wouldn't want to go there.
Anyway, this is all just speculation. I have no idea what copyleft enforcement looks like if termination is anything except "automatic&permanent". I don't think such enforcement has ever been done in the history of the planet, frankly [0].
Richard Fontana wrote at 22:32 (EDT) on Monday:
Ah, the principle that "mere nonsubstantive differences in termination provisions cannot lead to GPL incompatibility" never got codified (and may never have been fully thought out anyway).
This may be a more complex topic for cross-copyleft and license compatibility theory. Considering termination clause compatibility without a deep study of the complexities of the lex mercatoria of Free Software license compatibility probably puts the proverbial cart before the proverbial horse.
[0] Again, for purposes of reasonable discussion, I'm ignoring enforcement whose goal is to sell a proprietary license. We could invite Mickos to the list to tell us what the best termination policy is for single-copyright-holder GPL-misrepresentation cash shake-downs, but I don't think that'll make a better copyleft-next.
On 07/31/2012 07:54 PM, Bradley M. Kuhn wrote:
I also somewhat get why the second number must be 30 days (i.e., shorter), because the copyright holder has to ramp up legal action *after* the termination is permanent, so waiting 30 days really means waiting 90 or even 180 to get a Court to issue an injunction, since the 30 day mark is merely your *filing* date.
That is interesting, but you seem to be inventing a post-hoc justification for the 30-day provision that has no connection to why it's actually there.
Historically what happened was: 1) FSF got rid of automatic termination and added an unconventional 60-day repose provision that was originally worded in a way that many found quite confusing 2) Some commercial vendors and users, who didn't particularly understand what the FSF was trying to achieve with the 60-day repose provision, urged the FSF to add a cure provision (a common element of commercial software licensing agreements) 3) The FSF added a 30-day cure provision that was sort of integrated with the 60-day repose provision though improving the language of the latter. The idea for a 30-day cure provision didn't come from the FSF, but I believe it was the FSF that insisted that it be limited to first-time violators. 4) Person(s) associated with GPL enforcement in Germany later raised concerns about the absence of any automatic termination language.
The end result was what we have today, which is automatic termination coupled with possibilities for "provisional" and permanent automatic reinstatement, all integrated with the 30-day cure and 60-day repose constructs.
In other words, let's not establish some myth that the FSF sat down and decided to draft the perfect logical alternative to GPLv2-style automatic termination and thereby came up with what we have today in GPLv3.
Anyway, as I rereview GPLv3§8, I'm again convinced all moving parts are necessary.
Sure, you can decide now that it all works out perfectly and any other result would have been suboptimal. Our friends at Mozilla certainly like the GPLv3 provision well enough. But if that's so, it is serendipitous. Me, while I can see good and interesting things in the GPLv3 termination provision, I just think it's way more complicated than a termination provision needs to be. And I know from experience that it's hard for some people to understand.
- Richard
Richard Fontana wrote at 22:01 (EDT) on Tuesday:
On 07/31/2012 07:54 PM, Bradley M. Kuhn wrote:
I also somewhat get why the second number must be 30 days (i.e., shorter), because the copyright holder has to ramp up legal action *after* the termination is permanent, so waiting 30 days really means waiting 90 or even 180 to get a Court to issue an injunction, since the 30 day mark is merely your *filing* date.
That is interesting, but you seem to be inventing a post-hoc justification for the 30-day provision that has no connection to why it's actually there.
You're right: I should be clear: I wasn't consulted on the GPLv3 Termination clause, so *any* reason I might come up with is of course post-hoc justification.
Having admitted that, *every* time I look at the GPLv3 Termination section, I like it more.
urged the FSF to add a cure provision
Had I been asked, I would have supported a self-cure provision, just because it's good policy. Having a self-cure provision could have the effect to avoid an "oh well, I'm screwed now, why bother?" response by violators. The hope would be that they seek to cure in the timeframe allotted. I like it as a policy.
- Person(s) associated with GPL enforcement in Germany later raised
concerns about the absence of any automatic termination language.
I didn't know that. Is there a record of this and their concerns?
In other words, let's not establish some myth that the FSF sat down and decided to draft the perfect logical alternative to GPLv2-style automatic termination and thereby came up with what we have today in GPLv3.
I didn't mean to put forward the myth. *However*, I state again: when I look at what we have in GPLv3§8, I like what I see and think it's good policy.
But if that's so, it is serendipitous.
Agreed completely!
I just think it's way more complicated than a termination provision needs to be.
Ok, pursuant to my Platonic ideal thread, please describe what you see when you fly above the the Land of True Forms as a the One True Copyleft Termination Provision. If I tried to describe mine, I'd probably get close to GPLv3, but that may merely mean I have no imagination. :)
On 08/01/2012 06:08 PM, Bradley M. Kuhn wrote:
Richard Fontana wrote at 22:01 (EDT) on Tuesday:
- Person(s) associated with GPL enforcement in Germany later raised
concerns about the absence of any automatic termination language.
I didn't know that. Is there a record of this and their concerns?
The only recorded public record is in the GPLv3 Final Draft Rationale: http://www.gnu.org/licenses/gpl3-final-rationale.pdf (pdf, page 15, footnote 8). You can get some sense of when this change was made by considering the date of the "Last Call Draft".
Incidentally I recently deleted this whole complicated automatic termination, automatic reinstatement structure. European lawyers comfortable with the ex-Harvey Birdman Rule are invited to explain why it is more valuable than costly. In part the problem here is that there is so little on the record -- just that terse footnote.
I'll also point out that this last-minute, nonpublicized change did not result in greater acceptance of or advocacy for GPLv3 by the persons who requested it. This is not in itself a good reason to get rid of the last-minute change, but I'm not exactly shedding a tear in doing so.
I just think it's way more complicated than a termination provision needs to be.
Ok, pursuant to my Platonic ideal thread, please describe what you see when you fly above the the Land of True Forms as a the One True Copyleft Termination Provision. If I tried to describe mine, I'd probably get close to GPLv3, but that may merely mean I have no imagination. :)
Of all the provisions in a copyleft license a termination provision is the one that would seem to be least suited to Platonism. Practical considerations on the European side were what led to that last-minute change noted above. Given all your experience in GPL enforcement on the US side, would you *really* have come up with GPLv3 section 8?
- Richard
Richard Fontana wrote on 2 August:
Incidentally I recently deleted this whole complicated automatic termination, automatic reinstatement structure.
I thought about this, and I think I'm now comfortable with this "stricter" termination provision, provided my merge request below is also accepted.
I'll also point out that this last-minute, nonpublicized change did not result in greater acceptance of or advocacy for GPLv3 by the persons who requested it.
Agreed. While I think the notice/cure system designed for GPLv3 is fair and reasonable for violators, I certainly don't think these violators necessarily deserve the kindness, and the fact that violating companies still seem to prefer GPLv2 over GPLv3 indicates that they termination provision improvements of GPLv3 don't lead to wider adoption (apparently).
However, I think it's important that it's clear that the termination is permanent, and that the reinstatement procedure be contemplated in the text. I've submitted this merge request that does these things:
https://gitorious.org/copyleft-next/copyleft-next/merge_requests/20
On 10/12/2012 09:01 AM, Bradley M. Kuhn wrote:
However, I think it's important that it's clear that the termination is permanent, and that the reinstatement procedure be contemplated in the text. I've submitted this merge request that does these things:
https://gitorious.org/copyleft-next/copyleft-next/merge_requests/20
It seems obvious that termination will be permanent, and that any terminated license can be reinstated in some way by the licensor. Current copyleft-next does give a 30-day cure opportunity.
I'm inclined to reject this merge request, but perhaps I've misunderstood the rationale for it.
- RF
copyleft-next@lists.fedorahosted.org
-
Ben Cotton -
Bradley M. Kuhn -
Luis Villa -
Richard Fontana -
Stephen Michael Kellat