Should repo uploaders be allowed to delete their contributions at will?

[Znirk]

If my own code is GPL then I have the right to revoke that license and delete it.

Aaaactually, wrong: You can't revoke the GPL. Quoting section 2 of the GPL (version 3): "All rights granted under this License are granted for the term of copyright on the Program, and are irrevocable provided the stated conditions are met."


So you can stop distributing your code under the GPL, and you can start distributing it under a different license as long as it's all your own work. But the GPL remains valid for anyone who received a copy of your software under GPL and never violated its conditions, and those people have the right to redistribute the code under the same terms.

Which is exactly what I said. Do keep up.


D.

Wrong again It's probably exactly what you meant; but what you said seems misunderstandable.


The part I did misunderstand is the one where you write "I can't control copies that others may have downloaded". This sounds a bit like "I can't prevent people who have a copy from doing things they're not supposed to", when the real situation is "people who already have a GPL'd copy, and people who receive further copies from those people within the GPL's terms, et cetera perge ad infinitum, continue to have the right to modify and redistribute my code under certain conditions, and there's nothing I can do about that".


And just to try for some clarity from my side: My aim is not primarily to argue with ZXDunny, who just confirmed that we already agree on this, but to be clear towards other (potentially confused) readers about what options are and aren't legally possible in an I-no-longer-love-the-GPL situation.

 

[moxie]

I do not believe in making people be free and open by force. There is something inherently wrong in that thought. So, yes, a dev should most certainly be allowed to remove their work, should they wish to. In compliance with licenses and knowing that you cannot stop others from redistributing work under certain licenses, and so on and so forth, of course.

 

[ZXDunny]

And just to try for some clarity from my side: My aim is not primarily to argue with ZXDunny, who just confirmed that we already agree on this, but to be clear towards other (potentially confused) readers about what options are and aren't legally possible in an I-no-longer-love-the-GPL situation.

Glad we're all singing from the same hymn sheet. FWIW, my other project (PandaBAS, which is 100% written by me) is GPL, and the source is available for all via an SVN repository. Should I decide to no longer support it - and that will happen eventually - I will delete the SVN and the binaries I host. If there are no other backups of that source made by anyone else, it's unlikely that anyone would then be able to access the source code. To my knowledge, nobody else but me has a complete copy of the source; at least the SVN logs only show myself as being the single address that downloads from there.


I'd be interested to know where I stand if someone downloads a binary in January, I delete the source and the binary hosting in February and then that person demands the source, as per the GPL rules, in March.


D.

 

[skeezix]

A devs work is his own, he shoudl be able to delete it at will; the license, however, dictates distribution.


ie: The dev shoudl be free to delete it, period; if its a distributable license, then someone else can update it if he wants. End of story.


(of the 20+ stores I've had my software in, its always been this way)


jeff

 

[sswam]

there's sites like sam's alternative repo where stuff gets mirrored anyway. Proven stubborn, I bet PNDs taken down on the repo will remain there and possibly in other places, too.

Yup! at least that's how it's working for now

maybe this discussion only applies to those repo entries that allow redistribution. This should be the vast majority of the repo entries though.

Yes.

"The repo" needs an author's permission to redistribute that author's work.

Yes however once a PND has been released as free software, this permission to share and hack can't be revoked again for that PND release. So it would not make much sense for an unhappy maintainer to delete a free software PND, someone else will just upload it again; but this is inefficient and we might lose the comments, ratings, etc. So perhaps it's better to disallow or strongly discourage random deletes.


It would make more sense for an unhappy or busy maintainer just to quit maintaining the free-software PND, "orphaning" it. It can then be "adopted" by a different maintainer, it would be pointless to delete it and re-up it again. Free software PNDs may have a main developer, but they do not really have an "owner"; they have been given to and shared with the community. I know the dev may keep copyright, but that does not prevent other people maintaining, forking and developing the project.


Sometimes we might want to delete or archive old PNDs, if they are obsolete and cluttering up the repo or whatever. Such an action should be at least "seconded". And we should use a "fake delete" where the stuff is just moved into an archive for at least a few months or so.

I do not believe in making people be free and open by force.

I agree, that's why I prefer public domain (or MIT, BSD), but not the GPL or other copyleft! The GPL forces people to "share alike" but I reckon it's better to give them a choice, and allow the code to be reused by anyone in any project. GPL has random conflicts, even with other copyleft licenses. So for example you can't release a product based on code from Apache and GCC combined.

 

[_wb_]

I'd be interested to know where I stand if someone downloads a binary in January, I delete the source and the binary hosting in February and then that person demands the source, as per the GPL rules, in March.

AFAIK, when you offered it for download in January and someone downloaded it, you gave him a license that you cannot revoke (doesn't matter if he downloaded source only, binary only, or both). If he asks for the sources in March, you should still give them. The obligation remains for at least 3 years. You can avoid having this obligation if you don't distribute binary-only versions for download. See section 6 of the GPL.

 

[sswam]

FWIW, my other project (PandaBAS, which is 100% written by me) is GPL, and the source is available for all via an SVN repository. Should I decide to no longer support it - and that will happen eventually - I will delete the SVN and the binaries I host.

This is why we need to collect source code for all the free software PNDs that have been released. I'm trying to do that but it will take a while. Is your PandaBAS source / svn still up there, I don't see your svn folder there. Don't tell me you deleted it already? If you do still have the source, would you kindly provide me with a copy of it? thanks. It looks very good and it would be sad to lose it.

 

[ZXDunny]

I'd be interested to know where I stand if someone downloads a binary in January, I delete the source and the binary hosting in February and then that person demands the source, as per the GPL rules, in March.

AFAIK, when you offered it for download in January and someone downloaded it, you gave him a license that you cannot revoke (doesn't matter if he downloaded source only, binary only, or both). If he asks for the sources in March, you should still give them. The obligation remains for at least 3 years. You can avoid having this obligation if you don't distribute binary-only versions for download. See section 6 of the GPL.

And if I delete my binary and the only source of my code? I mean, I did offer source in the form of an SVN repository - it's not my fault the guy didn't download it when it was available.

FWIW, my other project (PandaBAS, which is 100% written by me) is GPL, and the source is available for all via an SVN repository. Should I decide to no longer support it - and that will happen eventually - I will delete the SVN and the binaries I host.

This is why we need to collect source code for all the free software PNDs that have been released. I'm trying to do that but it will take a while. Is your PandaBAS source / svn still up there, I don't see your svn folder there. Don't tell me you deleted it already? If you do still have the source, would you kindly provide me with a copy of it? thanks. It looks very good and it would be sad to lose it.

Don't worry, I still have the source freely available:


svn://alistairsserver.no-ip.org/specbas/


But be warned that although the PND gets sporadic updates, I update the source code almost daily. There are no sources for Forget-me-not and Spheres Of Chaos - they're not free, and the decision to distribute the source code lies with the original authors of those games; neither have agreed to any free distribution to my knowledge.


D.

 

[Znirk]

I'd be interested to know where I stand if someone downloads a binary in January, I delete the source and the binary hosting in February and then that person demands the source, as per the GPL rules, in March.

That's an interesting one. My understanding (but IANAL and feel like I'm beginning to get out of my depth here) is that you could do that without getting into legal trouble, but only because all the code is your own work. See http://www.gnu.org/l...eveloperViolate for the FSF's statement on not-really-this-exact-question.


The GPL is a standard format of you granting other people permission (or, to use a different Latin word, license) to do certain things with the code you wrote. You as the author don't need anyone's permission to do those things. In other words, when the GPL details what "you" can or must or mustn't do, then that doesn't mean you-the-author: it means you-the-user.


If you had used anybody else's GPL'd code in your work, you would be the user of those parts and have to supply source according to section 6 of the GPL at least for those components by other authors. If it's all your own work, though, you can simply take your ball and go home. It's a dick move, but it's legal.


If your code is then lost because of this, then yes, that's your users' own fault for not taking the GPL seriously enough to go get the source when it was available.

 

[skeezix]

The GPL is not as hard as people make it out to be


If _you_ offer something as GPL, you can do whatever you want with it, its _yours_; what the GPL is saying is if someone _else_ makes a derivative copy of it, then _they_ must also make their source available; even if they don't make it available right away, the GPL v2 (for example) just says, on request, they have to make reasomnable effort to get that source out. (ie: you should allow them some courtesy time, and they can send you a backup tape


Now, an interesting question is.. author posts source+binary; he can kill them and change the license as he sees fit (but not impacting the license of the one already released, even if identical, of course.) Buddy B comes along and modifies it, and posts his binary and source, well and good. 1 year goes by, and buddyB deletes his binary and source, but in the mean time C and D have sucked the binary down.


At this point..


B) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or,


Among other options; you see it is pretty specific.


Now it gets more interesting when you start releasing little patches; technically, you need to offer source for _all of them_, since any veriosn could have been downloaded; it can be tedious to do so.. so in general, most GPL projects just have a CVS, svn, git, etc up somewhere, so that folksw cna pull whatever version/diff/patch they may want.


Now, if you put up the source/binary and 2 years later, the serve is destroyed through no fault of your own, well, no one is really going to come bust your ass; you did your best, you took precautions to fulfill your part of the GPL, and shit happened.


jeff

 

[Edglex]

I dont think an idea like this is legally possible for the existing software, you'd need to get people to agree to some kind of license or terms and conditions containing these restrictions. And I wouldnt agree with it if it was. It's like using a service like google docs to make documents and then finding out later that google has claimed rights to all the content of your documents (something people have had concerns about). I certainly would not upload to the repo if it was going to cause me to lose rights to my work, which is effectively what has been suggested. Sorry, I see your point, and no offence meant, but I'm afraid I strongly disagree. The degree of control they get is what makes developers feel safe uploading to the repo, even to the point of commercial developers feeling comfortable with it now. If you take that away it will upset a lot of people and could even result in legal repercussions. I would also be against looking through the repo for stuff thats GPL and taking the developers repo deletion control away. They should be able to take it down if they decide to. Sure, someone else can put it straight back up, but they have control of their uploads. If a developer upsets people by taking away his work, unfortunately thats his/her right. It would probably annoy people but its the developers work to give and take as they decide.


Come on, the whole idea of the Pandora is about freedom!

 

[Caine]

Developers retain copyright to the applications they post. If, and only if, the license of the application allows it then somebody else can re-upload the application.


If you ever do re-upload an application which was removed by a developer then please take some effort to investigate exactly why he removed it. Most developers won't do this without a valid reason.

 

[_wb_]

Well, out of the four cases I'm aware of in which a repo maintainer decided to delete something, 2.5 were without valid reason:


- Streak has deleted stuff in a ragequit (not a valid reason)


- ZX_Dunny has deleted stuff because DREDD mirrored it (imo not a valid reason, but let's count this as 0.5)


- sebt3 has deleted something because he wanted to upload a newer version under a different name (let's count this as a valid reason, although using the update mechanism would have been a better solution imo)


- zLoud has deleted something, I'm not exactly sure why but it has something to do with the DREDD mirroring drama (not a valid reason)


Of course the Pandora is all about freedom. The question is: what freedom? The freedom of devs to fully control their packages, including deletion at any point in time without a valid reason? Or the freedom of users to download repo PNDs whenever they want, not in some arbitrary time window in which the dev is in a good mood.


I don't want to discourage devs who want to keep the full power to retract their contributions at any time. They can even use the repo, provided they don't tick the "redistributable" tickbox. But those devs that agree with redistribution can't consistently disagree with mirroring and archiving. For those PNDs, I think the repo should act as an archive / public backup service.

 

[Binky]

I don't want to discourage devs who want to keep the full power to retract their contributions at any time. They can even use the repo, provided they don't tick the "redistributable" tickbox. But those devs that agree with redistribution can't consistently disagree with mirroring and archiving. For those PNDs, I think the repo should act as an archive / public backup service.

quote++

 

[sebt3]

- sebt3 has deleted something because he wanted to upload a newer version under a different name (let's count this as a valid reason, although using the update mechanism would have been a better solution imo)

I was changing the appid which the repo couldnt help...

 

[_wb_]

Why did you need to change the appid? Nobody ever sees that anyway - it's not the same as the appdata dirname.

 

[milkshake]

well it could be that it had a generic id which had nothing to do with the PND, something that maybe is generated automatically by whatever tool sebt3 uses to compile and package his PND's in this case he would have needed to change it at somepoint anyway as its not ideal.

 

[Caine]

Of course the Pandora is all about freedom. The question is: what freedom? The freedom of devs to fully control their packages, including deletion at any point in time without a valid reason? Or the freedom of users to download repo PNDs whenever they want, not in some arbitrary time window in which the dev is in a good mood.

Freedom to fully control distribution of ones copyrighted work cannot be taken away from the copyright holder if he does not wish to give it away. You can create a repository which limits posted applications to those that protect freedom of users but that will only lead to multiple repositories.


In that regard, it is the freedom of the developers which matters here. Naturally users benefit from this as well (minor incidents excluded).


It seems a bit of a theoretical discussion though. In my opinion, there are few enough incidents to warrant any change to the repo.

 

[sebt3]

Why did you need to change the appid? Nobody ever sees that anyway - it's not the same as the appdata dirname.

this :

well it could be that it had a generic id which had nothing to do with the PND, something that maybe is generated automatically by whatever tool sebt3 uses to compile and package his PND's in this case he would have needed to change it at somepoint anyway as its not ideal.

Beside, please stop trying to show there is no point allowing dev to drop a PND. The law say the copyright holder can choose its distribution mean. That's it.

 

[Edglex]

Users are lucky that developers freely distribute work that they have spent a large amount of time on. It is the developers not the users work. If the developer wants to take it away, for whatever reason, that is up to them, full stop.


To me it is the same as if you owned a park or something or a house and let people come and look around or relax there. If you decide to not let them do so anymore, even if its because something just pisses you off or you had a bad day, they can't turn around and stop you from doing it. It's your house, in the same way that the work is the developers work. I dont care why people have taken their work down from the repo before, its theirs, so whether you agree with it or not the decision is also theirs.


Unless the license grants them some rights the users are out of the question here. Even in those cases the developer still holds copyright and still has rights to remove somthing even if someone else can put it back.


A step like this would make the repo a far worse, and far less free place in some respects than e.g. the apple store.That would be a very sad thing for the _open_ pandora.


I think that if this were implemented it would kill the repo almost instantly and probably also kill a lot of potential for future commercial development.


The repo is awesome as it is now - it does exactly what we want and pays attention to the needs of developers and users alike. Developers feel safe releasing software there. The system works - lets not mess with!