Mats Lidell writes: It's way worse than that. There is no separate named license for our documents. The license *is* the permissions notice, and it is, unfortunately, copyleft. Here's the relevant section: """ Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies. ＠ignore Permission is granted to process this file through TeX and print the results, provided the printed document carries copying permission notice identical to this one except for the removal of this paragraph (this paragraph not being relevant to the printed manual). ＠end ignore Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one. Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that this permission notice may be stated in a translation approved by the Foundation. Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided also that the section entitled ``GNU General Public License'' is included exactly as in the original, and provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one. Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that the section entitled ``GNU General Public License'' may be included in a translation approved by the Free Software Foundation instead of in the original English. """ It's not a problem for distribution. However, there is absolutely no other license that is compatible with this one in either direction, so syncing to other works (GNU, docstrings in XEmacs, etc) can only be done with explicit permission of the author of the content we want to add to our Texinfo manuals. There is no solution, without relicensing. We'll have to get a very good lawyer to do that. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Mats Lidell writes: That's right. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

On Wed, Apr 22, 2009 at 10:50 AM, Andreas Roehler <andreas.roehler(a)online.de&gt; wrote: If no license is specified, how do you know it is free software? Also, if an author specifies a specific set of rights to be granted, we cannot grant additional rights without the permission of the author, or else we have indeed "touched" the rights of that author. So, no, we can't just relicense everything under the GPLv3 unless we have been granted all pertinent rights by the authors. Sorry to have not been following up on my own list of issues to check. Real Life has been a bear for a couple of years now (although it's largely self-inflicted, so I can't complain too much). Can we put up a canonical copy of that list somewhere, so whenever someone has a little free time, they can grab an item and check it out? -- Jerry James http://loganjerry.googlepages.com/ _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas> I mean you can savely assume all code written for (X)Emacs Andreas> and sent in is meant to be distributed with it, is meant to Andreas> be with a license up to date. Anything else would bring the Andreas> work of the author simply out of use, which probably is not Andreas> the intent. In the worst case the author can't be contacted so the intent can't be checked. This implies we must assume the worst case which is that the intent was only to distribute the code as covered by the actual copyright. Andreas> Otherwise, should some author arg against publishing under GPLv3, Andreas> its time to remove that code still. Yes. And that is what we must do. We must identify the parts that is problematic and se what can be done about it. I'm trying to go down that path but unfortunately very slow. Hardly noticable I'm afraid. Andreas> Do you imagine an example, how someone might object GPLv3 Andreas> with the consequences seeing removed its code? Yes. Andreas> BTW what about to make an announcement at the XEmacs Andreas> html-site, telling authors about the license-change and Andreas> asking them to contradict if not agreed, assuming consent Andreas> otherwise? It will not stick. Real life can work that way but not in the world of copyrights. I think that some of the code hasn't even got any real author. It has just a copyright that doesn't allow us to relicens it as GPLv3 or later. We will not succeed in finding anyone that could grant us that right. So these files needs to go or be replaced. (I'm not a lawyer...) Yours -- %% Mats _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

SL Baur writes: But you have no paper to prove it. Your testimony would count in court, but I doubt it would stand up to a Darryl McBride waving an employment contract that one of our contributors had signed. Nope. You can attack that on two grounds. First, the FSF claims it *is* a derived work, but that depends on the legal theory that dynamic linking infects the source code which does not contain a copy of any GPLed work. Many lawyers disagree, and it has *never* been tested in court. Personally, I think if it were, the GPL would become redundant, because the court would decide that legally it means the same thing as the LGPL. But that's just an opinion. Second, the GPL quite clearly disclaims any such thing as an automatic GPLing of a work. If you distribute a GPL derivative without GPLing it, then you lose all your GPL rights to the original. But your work does not become GPL unless you say so, with the consequence that if you don't, both you and all your downstream must stop redistributing those parts of the work that are yours. Definitely not if you're not a lawyer. Me, I make assumptions safely when proving theorems. I don't consider any other assumption safe (especially not the assumption that Xlib calls documented by X.org as non-blocking won't block, but that's another rant). Yes, but it doesn't matter because the manuals are licensed under nameless (and therefore *not* upgradable) predecessor to the God- Forsaken Documentation License. BTW, Larry the lawyer says it's trivial to revoke assign.future if you want to. It might even be possible to revoke the assignment of past code on the basis that you never got your dollar (at least, I never got mine, and they didn't even ask if I cared), but the FSF would surely fight in court; it would be very expensive. I don't think that's worth it. If that's necessary we may as well just nuke it and start from scratch. But if we're going to start over, I'd prefer to use a system like doxygen or whatever and blow off the FSF manuals, instead working from the FSF docstrings (and GPLing our manuals in the process). It shouldn't be necessary, though. Ben has not been in touch for years, but he has always been happy to relicense to anything that would be useful. He also offered to assign all his documentation stuff to the FSF if they would let us use GFDLed Emacs docs under the nameless license. Stallman was a mild jerk about it (ie, he reiterated his usual demand that the FSF review every line proposed for relicensing), Ben got huffy, and then it was a contest to see who could be the bigger jerk until Stallman said "No docs for XEmacs or anybody who looks like you!" <sigh /> _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

SL Baur writes: Yes. My point is precisely that even that a**h*** could have won if he'd had the right docs. Unfortunately for him, they didn't exist. So am I. If it's code written from scratch, then either (1) it is just input to be processed by a GPL program at runtime, and the GPL makes it very clear that in that case there are no obligations on the distributor of the "input", or (2) it is derivative of the functions in (X)Emacs that it calls. Since those functions (ie, the implementing code) are not in the code written from scratch, they must be invoked, not copied, at runtime, and thus this claim depends on the dynamic linking theory. No, because it would be very hard to write such a thing without including byte-compiled expansions of macros, and those expansions do contain copyrighted XEmacs code. On the other hand, I think you could get away with distributing the .els under a proprietary license, at least if the courts enforce the law as written. None of the non-FSF lawyers I've talked to (including Larry Lessig and Larry Rosen) agree with you (at least not at the time I talked with them last, which in Lessig's case was about 4 years ago; he'd also had two glasses of wine by then). They are derived works only if they contain XEmacs code *or* are distributed as part of XEmacs. Referring to XEmacs code (eg, by referencing variables or invoking functions or subrs) is not "containing XEmacs code". Note that XEmacs, because it is free software, does not have the out of saying "these interfaces are internal, and we do not license their use". Thus all XEmacs interfaces are public. Calling a public interface has never been considered "derivation", because calling an interface is not copyrightable---if there's only one way to do it, it's not a matter of expression. I wouldn't be willing to try this in most cases, though, because there's a way in which *lots* of "expressive" code is almost sure to end up in your .so, and once again that's *macros*. I don't see why you think you'd need to. The GPL itself states those things explicitly (although not all in the same context). There's a big difference between relicensing under a very similar license, and assigning your copyright. Are you sure Ben wasn't resisting the latter? Sure. But IIRC, the FSF released 20.1 with its docs under the GFDL in 1998 or so, and for sure 21.1 was released with docs under GFDL in Oct. 2001. I hardly think that's useful. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas Roehler writes: Huh? Licensing is *the* right that is reserved to the authors. If no license is specified, then all rights are reserved to the copyright holders. You can distribute, betting on the good will of the copyright holders, but you *will* lose if they sue. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas Roehler writes: No, I'm sorry, that is no help legally, or ethically. There are many many people who have put one license on one file and another on a different one, for whatever reason. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas Roehler writes: I don't understand what example you want. Under U.S. law, it's absolutely clear. If there is no license, all rights are reserved to the rightsholder. Now, a license, like a contract, may be implicitly or verbally granted, but if you don't have written evidence it's your word against theirs, and the author will surely win in that case. If there is an explicitly written license, but it's not GPLv3- compatible, then you are out of luck. Again, there may be implicit or verbal grant of permission to upgrade, but you would lose in court if the rightsholder chose to withdraw it. I really don't think it is *ethically* right to *assume* permission, either. There are any number of people who have deliberately chosen, for one reason or another, to license their code as GPLv2 (or even GPLv1!) *only*. In any case where the permissions granted to us are unclear, we should make serious efforts to determine who owns the code or content, and get their permission. If that turns out to be impossible (at least two of our significant contributors have already left this vale of tears, for example), there might be grounds for doing something unilateral. But the only thing that protects our code from misuse[1] is copyright law. It is important that we respect it. Footnotes: [1] According to many of our contributors; me, I'd be happier with a permissive license. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

On Fri, 24 Apr 2009 21:07:31 +0200 Andreas Roehler <andreas.roehler(a)online.de&gt; wrote: No, you just remove the "or any later version". Then it is locked to the specific version of the GPL. Cheers, Sean _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Mats Lidell wrote: Hi Mats, probably I should come the an end with this matter. :) I'm afraid FSF software movement is stepping into a legal trap to the extend, it can't be called free anymore. That already started with famous `copyleft' notion, and developers considering GPL non-free have their point IMHO. Nonetheless I'm willing to weight the practical side - as long as (X)Emacs is available, thats my prove in first instance and as long I'm determined to contribute. Andreas PS Copyright notion is a trap in programming. Programming is about ideas, not line-counting copyright. The author of e=m*c**2 would have lost his case, because no judge would consider such a mince notation worthwhile in the sense of copyright. BTW the same in literature. Emily Dickinson never got paid for her poems because of copyright. Copyright is for the dull. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Marcus Harnisch wrote: You are right. However a sad example. Suppose its infected by famous `copyleft' rapist-towards-freedom story. Coercing freedom is a strange idea, thats a profound mistake. Can't assume RMS got it by himself. That not logical for me, as the file published with an older license will not vanish with a newer. Lets avoid that nightmare of negotiations. Lets accept Emacs as a present and lets send presents back. Otherwise I see rather a new hell than freedom coming. Andreas _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

SL Baur wrote: Hi Steve, I'm not sure grasping your words correctly. So please stay patient, should I respond something inappropriate in your eyes or if I'm not answering (i.e. being not sure about meaning) Just looking how to do something useful in a crazy world. Without knowing you: this world might be much more crazy you imagine. Here my personal story choosing the tools: The first book falling into my hands praised XEmacs the more advanced, more rich in features. GNU, however, was depicted slightly more stable. Then looking into the net, found flames against GNU Emacs, not the opposite. My decision after this was: GNU Emacs. Also let me remark: the amount of expressed feelings against RMS at this list is not met at emacs-devel vice versa. Don't remember anything against you, Ben Wing or Stephen Turnbull. Even not against XEmacs itself. RMS might have qualified XEmacs as from hell at some speech; even that sounded rather as a joke in my ears. Probably I miss important things here, never studied the issue; just delivering my impressions. We are able to peace as we are able to war. Let's choose peace. Andreas _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

SL Baur wrote: OK, don't have that experience and don't question yours. Let me say in an abstract sense though: outstanding people are capable of outstanding errors. Assuming that pertains to more then one person in FS-world. Lets enjoy it nonetheless - be patient with yourself too. Andreas _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Stephen J. Turnbull wrote: OK. So lets look at the practical side: Wanted to check in beginning-of-defun-function Could do it into the existing hg-repo, as it was already looked after, GNU got it while under GPLv2. However, other stuff is out there likewise. Might it be possible to switch all files with "or later" remark to v3 now? Thanks all Andreas it is not something you can disprove with logic. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Stephen J. Turnbull wrote: Thank you spelling out that so clearly. IMHO thats not just with GPL, thats the way our western politics treats people. It is not designed to make So far it fits well into the real world, however opposed to freedom, as legalizing misses all notion of it. Expensive paperwork and AFAIU discussed restrictions of GPL only take effect when distributing or linking code. Not, if you keep v2 and v3 code side by side in an repository. So can we make up a purely developing-repo, where the work with v3 might be done as far as possible? I could try to set up a git-clone with that purpose, if the hg-repo shouldn't be the appropriate place for it. (I'm interested at an git-repo by itself, as at the long run it looks the most promising.... However, yes, it needs some days of study, whilst hg and bazaar are considerably easier to start with. OTOH two or three days shouldn't count, should be justified with the background of a large project, delivering much more capabilities. And sure, bazaar is a great thing too already, whilst hg the most simple with this merit of its own for a larger community.) Andreas _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Stephen> We know what we need to do: make a list of the files with uncertain Stephen> permissions (this is basically done, but I forget where the list is), Stephen> contact their authors, if refused or no response then remove and if Stephen> necessary replace those files, If the files aren't necessary why not simply remove them? I think that that would be the first choice when going thought the list. Yes, and this is preferably done on the trunk on a daily basis. So the files needs to be brought up to a GPLV2 or later (or compatible) state before the big bang. Compatible here means either a copyright that can be relicensed as GPLV3 or later or can be kept as is eg isn't in the way. One example of the later would be the lispref-files. (If I understand it correctly.) So this is the plan I prefer: 1. Move the project into a state where we can perform a big bang GPLV3 or later shift. 2. Do the shift. 3. Live happily ever after ;-) Stephen> then do a global substitution on the permissions notice Stephen> throughout XEmacs, and commit that version. I'm working on a python script that does this. It replaces the known formulations of the GPLv2 or later copyrights with the GPLv3 or later. It also tries to categorise files according to what copyright they have so that for example the lispref files stand out. The idea I had was that we should run that script on the code base and it would spit out a list that would clearly state in what GPL-state each file is. Sort of "must-be-fixed", "ready-to-relicens", "ok-but-not-relicensable". What do you say? Waste of time or something to go for!? Stephen> It's just a lot of drudgery. Yes. Stephen> [...] in any case there is a ton of work that needs to be Stephen> done that would be perfectly usable sync'ing from GPLv2 Stephen> Emacs. What do you think of the possibility that the solution to questionable files is to replace them with GPLv2 files from Emacs? I fear that finding authors to some files will be difficult and for those it might be best just to go straight for a replacement. Yours -- %% Mats _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas Roehler writes: But what good does that do if it doesn't run? Specifically, upload/ adapt without testing at all is kind of silly, isn't it? It just sounds like extra work to me. The point is that *you* can combine free software any way you like in the privacy of your own machine. If you want to use GPLv3 code in your personal version of XEmacs, there is no problem whatsoever. If you have GPLv3 code in your private hg repository, it doesn't matter; you can still sync *from* the GPLv2 XEmacs "official" repository to your private repository. You just can't distribute your version of XEmacs with your new code until it's fully GPLv3, specifically you can't push to the public XEmacs repository. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Andreas Roehler writes: But what good does that do? The "transition" code doesn't run, so the upload is useless to anybody who doesn't have a mixed repo like yours. It's arguable that you can't distribute this code anyway, since it's a derivative of both <= v2 and >= v3 code according to the dynamic linking theory. Really, everything you've proposed here is clearly a way to circumvent the intent of the standard permissions notice. The only way to handle it respecting all of the applicable licenses is to convert XEmacs's own license. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta

Mats Lidell writes: The English syntax is complicated, I guess, but that's my intended meaning. Of course if they're useful but optional, they can be put on a list to try to recover them later. That's right. Most of our Texinfo files are inherited from GNU Emacs 19 and mixed with lots of work later; in practice it would be quite difficult to relicense them at all, and I imagine that the FSF would only allow the currently recommended version of the GFDL (or later version, as usual). The thing about code files is that my experience is that people are generally pretty picky about licenses for their original code, and often adamantly refuse to assign copyright even for derived code--- but they're usually quite happy to accept small changes in the license for the good of the project. (Of course you know about Bob Weiner and GPLv1, and probably you know about the SSLeay license that was deliberately designed to annoy the FSF and RMS. But that kind of thing is very exceptional.) However, since the FSF is a major stakeholder in our docs, I don't know if we'd have an option of anything but GFDL. The work has to be done, IMO. It could be done partly by hand if that's easier, since I doubt that the script will need to be used by us more than once or twice (and that in a short period of time). But if you don't mind writing a more robust script, you might want to put it on SourceForge or something for the benefit of others. If the Linux project ever decides to use GPLv3, you might even end up with free stock in Red Hat. :-) It's unlikely that that can be done very easily in the case of "necessary" files. But as an optimization, it should be the first thing to check. A quick diffstat should give a pretty good idea. We don't have to wait for replies from authors if there seems to be a feasible replacement, of course. Finding authors can be a problem; if we really need their code, the Review Board can discuss (after getting real legal advice) the option of a unilateral license change. There are precedents for this; the Wikipedia did it, for example. There is also a white paper on the OSI site by Eric Raymond that suggests that it might even stand up in court if challenged. You'd lose the specific code from that author but there needn't be other penalties. Presumably that white paper was written with the (informal) advice of a real lawyer. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta