2018-02-01 01:23:20-!- mode/#gentoo-council [+v The_Paya] by ChanServ 2018-02-01 03:15:56-!- mode/#gentoo-council [+v Tommy[D]] by ChanServ 2018-02-01 09:19:20<@mgorny> rich0: i don't think you strictly need to sign the FLA to have copyright notice for combined work as the guy who ulm talked to suggested 2018-02-01 09:19:59<@mgorny> in that case we can still assume Gentoo holds majority over all ebuilds 2018-02-01 09:20:12<@mgorny> and care only whenever someone explicitly put his name in additional copyright statement 2018-02-01 11:38:39<@ulm> mgorny: also what the policy currently says is too rigid 2018-02-01 11:39:59<@ulm> take PMS for example, it is copyrighted by 4 authors: https://dev.gentoo.org/~ulm/pms/head/pms.html 2018-02-01 11:40:54<@ulm> and I don't see us changing that to spb and others, or to ciaranm and others 2018-02-01 11:42:01<@ulm> but of course, if the foundation says so, I can take it off gentoo infrastructure 2018-02-01 11:42:36<@ulm> (in which case I would step down from the project though) 2018-02-01 11:44:53<+kensington> ulm: if the proposed policy keeps being crazy you won't be the other one 2018-02-01 11:47:46<@ulm> same here, 4 authors: https://gitweb.gentoo.org/proj/ebuild-mode.git/tree/ebuild-mode.el 2018-02-01 11:48:47<@ulm> and what would we put there? mkennedy committed the original version, but by number of lines I would be the main author by now 2018-02-01 11:50:24<@ulm> so... 2018-02-01 11:50:56<@ulm> I think the policy should say that the copyright notice can list all authors 2018-02-01 11:51:47<@ulm> but a simplified form can be used, e.g. in case of ebuilds, listing only the main author "and others" 2018-02-01 12:10:35-!- mode/#gentoo-council [+v kentnl] by ChanServ 2018-02-01 12:11:23-!- mode/#gentoo-council [+v kensington_] by ChanServ 2018-02-01 12:15:49<@mgorny> ulm: well, i'd go with the following three points: 2018-02-01 12:16:26<@mgorny> 1. we encourage copyright main author (+ 'and others') 2018-02-01 12:16:50<@mgorny> 2. for projects where that makes sense, we encourage treating it as one big thingie rather than separate files 2018-02-01 12:16:56<@mgorny> (which simplifies 'main author') 2018-02-01 12:17:15<@mgorny> 3. if necessary by license requirements, extra copyright lines may be present 2018-02-01 12:59:34<@ulm> mgorny: and IMHO the gentoo repo is such a project 2018-02-01 12:59:48<@mgorny> i agree 2018-02-01 12:59:53<@ulm> because ebuilds don't exist indenpendently of each other 2018-02-01 13:00:05<@mgorny> and there's a lot of data being copied around 2018-02-01 13:00:21<@ulm> yep, and they rely on eclasses and profiles 2018-02-01 13:00:38<@mgorny> foo-2.ebuild is usually not a new file but a copy of foo-1.ebuild 2018-02-01 13:02:02<@ulm> we also treat it as one big project, e.g. commit access is for the whole repo, and we have a single license for it 2018-02-01 13:02:06<@ulm> IMHO 2018-02-01 13:04:28<@ulm> one bugzie product, too 2018-02-01 13:04:49<@ulm> (we used to have games separately, but that's no longer the case) 2018-02-01 13:39:13<+rich0> ulm: it would still be copyrighted by 4 authors. It just wouldn't put all four names in the notice 2018-02-01 13:40:16<+rich0> I'm not sure why we're even writing up a policy if it basically ends up saying "just do whatever you want in every individual project" 2018-02-01 13:40:34<+rich0> If it doesn't actually provide some kind of constraint on what people can do, then what stops another eudev mess? 2018-02-01 13:40:58<+rich0> By all means list the 4 authors in the file if it bothers people that much - just don't do it in the copyright notice 2018-02-01 13:41:38<+rich0> But, maybe Gentoo projects are just independent enough that trying to call them all "Gentoo" or unify them just doesn't make sense, in which case I'm not sure what value Gentoo is adding to the projects. 2018-02-01 13:41:58<+rich0> Is there a reason PMS has to be hosted on Gentoo infra, or portage, or even the ebuild repo? 2018-02-01 13:42:18<+rich0> If each team just does its own thing, they can just as easily do that if they're all independent... 2018-02-01 13:47:02<+rich0> Also, what happens if somebody wants to take an ebuild outside the gentoo repo that has a non-Gentoo copyright notice on it, which is GPL2+, and put it in the main repo? How does that work if you treat the whole tree as if it is one work? 2018-02-01 13:47:12<+rich0> Even the Linux kernel isn't treated as a single work for copyright in this way. 2018-02-01 13:47:25<+rich0> And I'd argue the kernel is far more unified a body of work than an ebuild. 2018-02-01 13:49:51<+rich0> kensington_: this is the first I'm hearing that the policy "is crazy" 2018-02-01 13:51:59<+rich0> What about it don't you like? 2018-02-01 13:55:21<+rich0> I suppose the mgorny's #1/3 might be possible as long as we try to steer people down the simplest reason. Note that mgorny's #3 contradicts #2 essentially. 2018-02-01 13:56:17<+kensington_> rich0: is there an up-to-date document? 2018-02-01 13:57:22<+rich0> kensington_: sounds like ulm has the most recent one 2018-02-01 13:57:34<+rich0> Though it doesn't contain the simplifications in the bug 2018-02-01 13:58:26<+rich0> Keep in mind that the point of a copyright notice isn't to give credit to the author. 2018-02-01 13:58:33<+rich0> It is to warn the reader that the work is copyrighted. 2018-02-01 13:59:27<+rich0> If you contributed something to a work you have copyright on it whether your name is listed or not. 2018-02-01 13:59:42<+rich0> Unless you've assigned that copyright to somebody else, in jurisdictions that permit this. 2018-02-01 13:59:59<+kensington_> I saw some very 'interesting' ideas thrown around but I'll refrain from bikeshedding until I see something more recent 2018-02-01 14:04:28<@ulm> rich0: as I said, I would step down from the PMS project if the foundation would tell me that I must take my name off its copyright notice, or that it cannot be hosted on gentoo infra 2018-02-01 14:05:02<+rich0> ulm: well, having your name on the policy isn't inconsistent with the policy. :) 2018-02-01 14:05:07<+rich0> err, on PMS 2018-02-01 14:05:18<@ulm> rich0: that ebuild example is quite academic 2018-02-01 14:05:41<@ulm> in practice, we can simply refuse accepting such ebuilds in the tree 2018-02-01 14:05:43<+rich0> ulm: you haven't seen wltjr's java repo? 2018-02-01 14:06:03<@ulm> just rewrite the stuff from scratch, ebuilds aren't so complicated 2018-02-01 14:06:47<@ulm> and if you arrive at an almost identical ebuild, then it shows that it wasn't copyrightable, in the first place :) 2018-02-01 14:07:11<+rich0> ulm: sounds like a great argument to spend $50k making in court... :) 2018-02-01 14:07:36<@ulm> kensington_: thist should be the most recent draft: https://dev.gentoo.org/~ulm/glep-copyrightpolicy.html 2018-02-01 14:08:29<+rich0> ulm, kensington_: that doesn't contain the proposed changes under attribution, though obviously there still is some work to be done in cleaning that up beyond this 2018-02-01 14:08:47<@ulm> no it doesn't 2018-02-01 14:09:06<+rich0> Those are in: https://bugs.gentoo.org/642072#c4 - though I'm also in favor of getting rid of the (see below) stuff 2018-02-01 14:09:55<+rich0> If people are unwilling to participate in projects if their name isn't in the copyright notice I guess it won't kill us to let them insist on it, whether it be PMS, ebuilds, or whatever. 2018-02-01 14:10:03<@ulm> yes, drop the "see below" part 2018-02-01 14:11:00<@ulm> but I still think that it's too rigid 2018-02-01 14:11:47<@ulm> what is wrong with a notice like this? "© 2007–2018 Stephen Bennett, Christian Faulhammer, Ciaran McCreesh and Ulrich Müller. Contributions are owned by their respective authors [...]" 2018-02-01 14:12:08<+rich0> ulm: I'd probably add "and others" to that as well - I'm sure others contributed 2018-02-01 14:12:17<+rich0> I would ditch the last sentece 2018-02-01 14:12:58<+rich0> That is true whether we state it or not - the only reason to even have the line is to align with copyright law, which doesn't include that statement 2018-02-01 14:13:24<@ulm> yeah, that can be discussed, ""© 2007–2018 <4 authors>, and others." is fine with me 2018-02-01 14:14:01<@ulm> but I won't want to start a discussion with ciaranm about who's name has to be dropped 2018-02-01 14:14:37<+rich0> I think we should word the policy in a way that allows for this, but which doesn't suggest that in general we should keep appending names to the line 2018-02-01 14:14:56<+rich0> We really don't want everybody who makes a commit to a file to keep adding their names to the copyright notice 2018-02-01 14:15:52<+rich0> Legally they could, and of course somebody could ragequit if we didn't let them, but I think it is a bad practice in general. 2018-02-01 14:16:29<+rich0> The policy should still allow this though. 2018-02-01 14:16:36<+rich0> So, if somebody wants to add their name to an ebuild, that is fine 2018-02-01 14:16:41<+rich0> Just discouraged 2018-02-01 14:17:04<+rich0> We could publish annual stats on the length of the longest copyright notice. :) 2018-02-01 14:18:11<+rich0> Sorry - thinking out loud. I guess I'm struggling to think of how to word things to make it clear that it is fine to do for PMS, but not fine to do with every other random ebuild/project/etc, unless it is fine everywhere. 2018-02-01 14:19:01<+rich0> Legally I think we're ok either way. There is nothing wrong in law with having a 3000 char copyright notice. 2018-02-01 14:20:20<+rich0> And we could always drop names from notices after the authors have stepped down and are no longer in a position to threaten to quit if they haven't written more than half the document. 2018-02-01 14:20:30<+rich0> That would work for PMS for some of those names. 2018-02-01 14:20:42<+rich0> Again, copyright notices aren't credit. 2018-02-01 14:32:49<@dilfridge|mobile> rich0: ulm: for a (non-gentoo related) Perl package I wrote a release plugin that generates exactly this from git history... 2018-02-01 14:33:05<@dilfridge|mobile> (Per file) 2018-02-01 14:35:28<@ulm> rich0: how about this? https://paste.pound-python.org/show/rOYR1oHrnLgFbpC4z5Yu/ 2018-02-01 14:36:09<@ulm> hm, and strike the second "(see below)" 2018-02-01 14:37:42<@ulm> rich0: for PMS it's more complicated than that 2018-02-01 14:38:10<@ulm> because simple "git blame" or counting number of commits is very misleading there 2018-02-01 14:39:14<+NeddySeagoon> What is a line? Code? What happens when the copyright notice is >50% of the file ? 2018-02-01 14:39:46<@ulm> the copyright notice is not copyrightable and therefore not covered by the copyright notice :) 2018-02-01 14:39:54<@mgorny> ulm: you've removed 'see below' from the first snippet but not from second 2018-02-01 14:40:13<@ulm> mgorny: yes, that's why I said hm, and strike the second "(see below)" 2018-02-01 14:40:30<@mgorny> lol, i interpreted that literally ;-D 2018-02-01 14:40:36<@mgorny> too much work for today, i guess 2018-02-01 14:41:01<@ulm> mgorny: https://paste.pound-python.org/show/oFGkHvxF8JLEjjKtDo6J/ :) 2018-02-01 14:46:22<@ulm> also, what about the wiki? we don't have copyright notices on wiki pages 2018-02-01 14:47:15<+leio> you still going by the theory that copyright notices have some sort of actual value? 2018-02-01 14:47:37<+leio> my understand is that they are just a way to guard against misuse, more of an informational thing 2018-02-01 14:48:13<+leio> which is why I find claims like "I must have agreed to a copyright assignment to Gentoo Foundation because I seem to have been fine with the repoman forced copyright notice" that flew around rather outrageous 2018-02-01 14:48:36<+leio> I consider it more of a "it's a real entity, it's more deterrant than my name" when keeping it in there (and then repoman doesn't allow to add mine) 2018-02-01 14:48:52<+leio> though haven't followed the very latest discussions; also waiting for some more concrete intermediate result in writing 2018-02-01 14:49:06<+leio> real bigger entity* 2018-02-01 14:49:14<@ulm> leio: in US law there is a provision that without an explicit copyright notice it can be "innocent infringement" 2018-02-01 14:49:32<+NeddySeagoon> leio: It is. It allows the entity to act instead of getting to agree ... if you can find them. 2018-02-01 14:49:33<+leio> sure, that doesn't sound it matters what the actual holder is claimed in the notice 2018-02-01 14:49:43<+leio> that was to ulm 2018-02-01 14:49:49<+leio> I don't have a comment to neddy's claim right now 2018-02-01 14:50:11<@ulm> leio: I agree, it doesn't matter much 2018-02-01 14:50:27<@ulm> leio: but it shouldn't be completely wrong 2018-02-01 14:51:33<@ulm> I still think that it would much simplify things if we treated the gentoo repo as a single work 2018-02-01 14:51:34<+NeddySeagoon> leio: I agree with your "outragous" but I don't think it matters much as long as readers are aware that someone, somewhere holds the copyright. 2018-02-01 14:53:01<+NeddySeagoon> It avoids the "innocent infringement". 2018-02-01 14:56:18<@ulm> well, the problem is to find some balance between accuracy, brevity, and maintainability 2018-02-01 14:56:31<+leio> ulm: yes, but the problem is that people copy ebuilds to their overlays, etc. That won't have that information 2018-02-01 14:56:58<@ulm> putting more names into the line may make it more accurate momentarily 2018-02-01 14:57:14<@ulm> but less maintainable 2018-02-01 14:58:03<@ulm> so I think that for ebuilds we would want to have a single entity there, plus "and others" 2018-02-01 14:58:53<@ulm> for projects other than ebuild repo authorship will be easier to trace 2018-02-01 14:59:23<@ulm> so the balance point may be different, and list more than one copyright holder 2018-02-01 15:01:14<+leio> I also find most ebuilds to not be copyrightable work, but whatever 2018-02-01 15:01:34<+leio> that's for the courts to decide; but those that just list depends and such, questionable 2018-02-01 15:02:33<@ulm> leio: yes, but if we distinguish between copyrightable and non-copyrightable ebuilds, we'll end up adding and removing notices all the time 2018-02-01 15:02:50<@ulm> an ebuild may start as trivial and grow over time 2018-02-01 15:03:04<@ulm> or the other way around 2018-02-01 15:03:53<@ulm> and I think it won't harm to have a notice on trivial ebuilds too 2018-02-01 15:05:17<@ulm> again, it's simpler if the whole tree is the work, because that's certainly copyrightable 2018-02-01 15:07:24<@ulm> just add a line "This file is part of Gentoo Linux." 2018-02-01 15:07:43<+leio> and then go cease and desist lettering all overlays? ;p 2018-02-01 15:07:47<@ulm> or "GNU/Linux", so we have another thing to bikeshed :) 2018-02-01 15:07:47<+rich0> ulm: and what if you want to do a hostile fork of somebody else's overlay? 2018-02-01 15:08:20<+rich0> This whole thing was started by eudev... 2018-02-01 15:09:22<@ulm> rich0: that penultimate line of yours is about which comment of mine? 2018-02-01 15:09:35<@ulm> (sorry, I'm getting a bit confused) 2018-02-01 15:09:43<+rich0> Forgiveable :) 2018-02-01 15:09:59<+rich0> "again, it's simpler if the whole tree is the work, because that's certainly copyrightable" 2018-02-01 15:10:28<+rich0> Taking a stand that individual ebuilds aren't distinct works is fine until somebody puts some code in one, and somebody outside of Gentoo decides it is a distinct work 2018-02-01 15:11:35<@ulm> http://git.savannah.gnu.org/cgit/emacs.git/tree/admin/notes/copyright#n16 2018-02-01 15:11:51<@ulm> ^^ contains some useful guidelines about that aspect 2018-02-01 15:12:44<+rich0> ulm: sure, but keep in mind that they require assignment for all contributions 2018-02-01 15:13:21<+rich0> Still, we should read and consider if there is anything useful 2018-02-01 15:14:00-!- mode/#gentoo-council [+v kensington] by ChanServ 2018-02-01 15:16:56<@mgorny> K_F: i'm not sure if you should've mentioned the patch 2018-02-01 15:17:08<@mgorny> now they can claim they can't use it because i didn't sign their copyright crap 2018-02-01 15:17:18<@K_F> nah 2018-02-01 15:17:37<@K_F> thats no problem, and its easier for werner to deal with the patch 2018-02-01 15:17:40<@ulm> rich0: ok if I update the draft as in the pastebin? 2018-02-01 15:17:48<@K_F> will speak with him a bit tomorrow in brussels 2018-02-01 15:18:18<@ulm> rich0: i.e. with the sentence "Additional contributors can be listed, but this is neither required nor recommended." 2018-02-01 15:18:38<+rich0> ulm: sure - let's at least get a straw man out there and think about it 2018-02-01 15:19:55<@ulm> also "Abstract", "Motivation", and "Rationale" should be filled in, but I guess I'll wait with that until trustees say that it's ok to publish in glep format 2018-02-01 15:20:18-!- mode/#gentoo-council [+v alicef] by ChanServ 2018-02-01 15:20:32<@ulm> rich0: BTW, you don't happen to have an index of previous discussions? 2018-02-01 15:27:09<+rich0> Not as an index, but they should be on the archives 2018-02-01 15:30:47<@ulm> rich0: certainly they should be there, problem is to find them 2018-02-01 15:32:26<@ulm> and there are discussions on forums too, e.g. this: https://forums.gentoo.org/viewtopic-t-98610-highlight-copyright+assignment.html 2018-02-01 15:40:00<@ulm> rich0: this is an interesting post: https://archives.gentoo.org/gentoo-dev/message/60630a3e1b5ba40c49fa65daadd45fbd 2018-02-01 17:30:44-!- mode/#gentoo-council [+o whubbs] by ChanServ 2018-02-01 17:52:01-!- mode/#gentoo-council [+o mgorny] by ChanServ