| |||||||||||||
Launchpad source releasedLaunchpad source releasedPosted Jul 21, 2009 16:56 UTC (Tue) by agrover (subscriber, #55381)In reply to: Launchpad source released by davi Parent article: Launchpad source released
Well I guess it would be a nice rah-rah gesture, but unlike Launchpad I don't really see there being much value to the LWN codebase without the LWN content and editorial staff.
(Log in to post comments)
Launchpad source released Posted Jul 21, 2009 17:43 UTC (Tue) by lacostej (guest, #2760) [Link]
What about external contributions ?
Launchpad source released Posted Jul 21, 2009 19:30 UTC (Tue) by NAR (subscriber, #1313) [Link]
I've once read an Erlang/OTP developer about patches from external contributors. He wrote that virtually all patches they receive have to be rewritten basically from scratch. The patches break existing stuff (automatic tests are useful to catch this), are using an other coding style, don't contain documentation comments, etc. In the end it takes a lot of time and effort to manage these patches, nearly that much time it would take to develop the code internally from scratch.
I do think that the situation with the LWN engine would be very similar. I suppose the editors don't have much time to work on the engine, so they wouldn't have much time handling those external patches. So these patches would just fill a mailbox somewhere...
Launchpad source released Posted Jul 21, 2009 18:03 UTC (Tue) by davi (guest, #18853) [Link]
What about know the code which manage us? Is it collecting user statistics, etc.?
Launchpad source released Posted Jul 21, 2009 19:22 UTC (Tue) by NAR (subscriber, #1313) [Link]
How would you know that the distributed source has anything to do with the code running at lwn.net?
Launchpad source released Posted Jul 21, 2009 20:02 UTC (Tue) by davi (guest, #18853) [Link] If the LWN web application were licensed under Affero GPL version 3 or later the webapp itself would have to show a link to offer a download of the same source which is running the site at that time. See section 13, Remote Network Interaction of the Affero GPL v3. A cron job can pack the tar ball of the source code which running the web app at that time, as it is done for example at the GNU Herds web application
Launchpad source released Posted Jul 21, 2009 20:56 UTC (Tue) by jspaleta (subscriber, #50639) [Link]
Hmm, an exact copy of what is running. I wonder how Canonical plans to bring Launchpad.net into compliance with that requirement. Launchpad.net doesn't yet implement a homepage link to a tarball like gnuherds does.
-jef
Launchpad source released Posted Jul 21, 2009 21:12 UTC (Tue) by foom (subscriber, #14868) [Link]
Quite obviously, they don't need to: since they are the copyright owners, they don't need to follow
the AGPL's terms at all...
Launchpad source released Posted Jul 21, 2009 21:20 UTC (Tue) by jspaleta (subscriber, #50639) [Link]
Indeed, and as they require contributors to affirmatively assign copyrights over to Canonical before contributions can be accepted the AGPL won't be provide any contributed contributed code either. That's a nice little clever twist. If Canonical wants to run a modified Launchpad.net out of a private bzr branch that is not publicly available they will still be able to while an external instance of the code will have to figure out how to solve the running code distribution compliance problem on their own. Very clever.
-jef
Launchpad source released Posted Jul 21, 2009 21:44 UTC (Tue) by job (guest, #670) [Link]
You don't need the AGPL in order to do that. No need to bring in the conspiracies. Canonical is the copyright holders so they can run a private branch regardless of how they license the software to others.
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 3:30 UTC (Wed) by khim (subscriber, #9252) [Link] You don't need the AGPL in order to do that. What other license forces you to track actual changes and publish them? Even proprietary licenses are usually not so strict: they often limit number of instances you can run, but they certainly don't require for you to track what is installed where. Funny how in the name of openness we've lost one of the advantages of free software. I certainly will consider proprietary alternative before I'll think about AGPLed stuff. Sure, proprietary stuff is bad, but AGPL is worse.
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 4:33 UTC (Wed) by davi (guest, #18853) [Link]
> Funny how in the name of openness we've lost one of the advantages of free software. I certainly will consider proprietary alternative before I'll think about AGPLed stuff. Sure, proprietary stuff is bad, but AGPL is worse.
Free Software is about users' freedom, not about developer convenience. This is not just Open Source.
Free software is a matter of the users' freedom to run, copy, distribute, study, change and improve the software. Read http://www.gnu.org/philosophy/free-sw.html
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 11:06 UTC (Wed) by SEMW (guest, #52697) [Link] > Free Software is about users' freedom, not about developer convenience. This is not just Open> Source. Free software is a matter of the users' freedom to run, copy, distribute, study, change and > improve the software. You claim to contrast Free Software to Open Source, but the users' freedoms you list are freedoms common to both the FSF Free Software definition and the OSI Open Source definition. (In fact, apart from slightly different wordings, I don't see that there is any substantive difference between the two definitions in terms of the freedoms they guarantee. The only licence in Wikipedia's list that the OSI lists as complying with its definition but the FSF does not for its is the Artistic licence 1.0; and that only because the FSF think it's "too vague".)
Did so. Many years ago. Posted Jul 22, 2009 21:00 UTC (Wed) by khim (subscriber, #9252) [Link] Free software is a matter of the users' freedom to run, copy, distribute, study, change and improve the software. Read http://www.gnu.org/philosophy/free-sw.html If you actually read the thing and now blindly waved link around you'd know this part: You should also have the freedom to make modifications and use them privately in your own work or play, without even mentioning that they exist. If you do publish your changes, you should not be required to notify anyone in particular, or in any particular way. And were you somewhat observant you'd remember that years ago Stallman was adamant about this right (for example here). Sure rise of the Web and SaaS approa changed his position and now he considers AGPL free and even changed the GPLv3 to accomodate, but in my view AGPL was and is a non-free license, sorry. P.S. Oh - and it's strictly prohibited here, too. Just a matter of policy, we are not urhing anyone to boycott it, but we sure as hell will not be using any code released under this license - launchpad included.
Did so. Many years ago. Posted Jul 22, 2009 22:00 UTC (Wed) by Wol (guest, #4433) [Link]
You've missed a *key* word. "and use them *privately* in your own work and play". I don't think running a public web server can be classified as "private use".
And (forgive me if I'm wrong) I doubt the AGPL contains a requirement to notify anybody other than your users that you're using a modified version (and the ordinary GPL requires that, too, in the sense that anybody to whom you distribute a binary is a user).
So, I'm afraid I'd agree with RMS. And I don't think he's modified his stance, either. If somebody provides me with binaries based on GPL sources, they have to tell me that it's based on GPL code, and they have to make that code available to me. The AGPL is in principle exactly the same - if they provide a web service based on AGPL code, then they have to tell me that it's based on AGPL code, and they have to make that code available to me.
The only exception to this (which also applies equally to GPL and AGPL code), is that a company is considered "a user", so internal company use does not trigger the "must share the code" requirement, because an employee is not a user in the legal sense.
Cheers,
Did so. Many years ago. Posted Jul 22, 2009 23:00 UTC (Wed) by khim (subscriber, #9252) [Link] You've missed a *key* word. "and use them *privately* in your own work and play". I don't think running a public web server can be classified as "private use". Why not? Changes to the free programs installed by universities were kept private for years - students were certainly not entitled for a copy. What's the difference with web server? Heck, the very same site we are discussing this uses the free code in this manner! And (forgive me if I'm wrong) I doubt the AGPL contains a requirement to notify anybody other than your users that you're using a modified version (and the ordinary GPL requires that, too, in the sense that anybody to whom you distribute a binary is a user). If the code is in compiled language then creation and distribution of modified version is significant effort by itself. If you distributed code written in iterpreted language - the requirement was fullfilled automgically. Additional burden was negligible. AGPL turns what is trivial and easy process (add or remove few lines from code on server) to onerous and potentially dangerous process (not only you can accidently publish passwords but opens the door for automatic attacks). The AGPL is in principle exactly the same - if they provide a web service based on AGPL code, then they have to tell me that it's based on AGPL code, and they have to make that code available to me. No. There ara big difference. Let's take a look on the basic four freedoms: As you can guess there are times where these freedoms are in conflict. GPL resolves them in order: freedom 0 is the most important, freedom 3 is the least important. AGPL turns the whole paradigm on it's ear: freedom 2 is made more important then freedom 1 and freedom 3 is turned from option to requirement. If this is not a change of stance then what it is?
Did so. Many years ago. Posted Jul 24, 2009 14:36 UTC (Fri) by Los__D (guest, #15263) [Link] As you can guess there are times where these freedoms are in conflict. GPL resolves them in order: freedom 0 is the most important, freedom 3 is the least important. AGPL turns the whole paradigm on it's ear: freedom 2 is made more important then freedom 1 and freedom 3 is turned from option to requirement. If this is not a change of stance then what it is? What? The Free Software Definition mentions nothing about differing importance of each freedom, nor that freedom 3 is optional; A program is free software if users have all of these freedoms.
Did so. Many years ago. Posted Jul 25, 2009 5:58 UTC (Sat) by khim (subscriber, #9252) [Link] The Free Software Definition mentions nothing about differing importance of each freedom, How the conflicts are resolved then? If they all are equally important then the only free license is "public domain" - everything else limits freedom 3. nor that freedom 3 is optional. Sorry, but you have funny idea about what freedom is. If you are persona non-grata in Russia and so can not visit Siberia - your mobility freedom is affected, if you can only live in Siberia (and in particular place at that) - it's called banishment and hardly can be called "freedom" too. With GPL my freedom 3 is somewhat affected: if I decide to distribute binary - I must promise to give recepient sources too. It's fair: freedom 3 is only freedom 3, after all, and freedoms 0, 1, and 2 come first. With AGPL I have no choice: if I use the code in my web app - I my distribute both binaries and sources despite my protest - and this violates freedom 0. You and RMS can play fast and loose with definitions as long as you want, but I read them in order and this approach clearly says that AGPL is not free.
Did so. Many years ago. Posted Jul 25, 2009 8:19 UTC (Sat) by Los__D (guest, #15263) [Link] You are certainly entitled to your own interpretation of the four freedoms, but the FSD contradicts you. Sorry, but you have funny idea about what freedom is. This is not my idea about what freedom is, but the FSD's. I seem to have to repeat the quote: A program is free software if users have all of these freedoms. You and RMS can play fast and loose with definitions as long as you want, but I read them in order and this approach clearly says that AGPL is not free. It seems like you are the one doing the playing; the only way Freedom 3 would be affected, is if you only read the first part of it, and not the freedom as a whole, or if you add "under any license you want".
Argh. Posted Jul 27, 2009 1:55 UTC (Mon) by khim (subscriber, #9252) [Link] This is not my idea about what freedom is, but the FSD's. Huh? Where the hell freedom is defined in that essay? There are definition of free software and there are fundamental, basic "freedom" concept. The former is indeed defined in the FSF's article, the latter can be found in different dictionatries. For example wictionary says: FreedomWebster says: Free-dom: This is freedom. The absence of necessity, coercion, or constraint. It seems like you are the one doing the playing; the only way Freedom 3 would be affected, is if you only read the first part of it, and not the freedom as a whole, or if you add "under any license you want". No. There are no freedom with AGPL. I have duty to release changes. I am obliged to release changes. That's not freedom - that's exact opposite. When I'm forbidden from releasing changes (proprietary license) - that's not freedom, when I'm forced to release changes - that's not freedom, I only ever have freedom number 3 when license allows me to release changes or keep them private. GPL affects this freedom to, but only where it's needed to protect other, more important freedoms, while AGPL does this unconditionally.
Arghx2 Posted Jul 27, 2009 2:11 UTC (Mon) by kragil (subscriber, #34373) [Link]
Nonsense. Just unplug the computers network and you aren't obliged to release anything.
For computers without network devices GPL and AGPL are essentially the same and provide the same freedom. But I am certain you won't agree.
True, but only partically true... Posted Jul 27, 2009 16:03 UTC (Mon) by khim (subscriber, #9252) [Link] Nonsense. Just unplug the computers network and you aren't obliged to release anything. True. For computers without network devices GPL and AGPL are essentially the same and provide the same freedom. Not even close. But I am certain you won't agree. Of course not! If I unplug the computer from network I'm getting freedom 3 back but I lose freedom 0 - and freedom 0 is the most important freedom of all four! So while "just unplug the system from network" is a way to get freedom 3 back it's in no way makes AGPL free license: A program is free software if users have all of these freedoms - and this "unplugging act" loses the most important freedom.
Argh. Posted Jul 27, 2009 7:36 UTC (Mon) by Los__D (guest, #15263) [Link]
Nice to see that you are able to choose the one (of the several) definitions that agrees with you.
The AGPL and GPL is exactly the same in this regard, no matter which funny ways you decide to read the four freedoms.
Care to point any other definition? Posted Jul 27, 2009 16:35 UTC (Mon) by khim (subscriber, #9252) [Link] Nice to see that you are able to choose the one (of the several) definitions that agrees with you. Actually they all are pretty similar. "Absence of necessity, coercion, or constraint", "liberation from slavery or restraint or from the power of another", and other similar definitions all point to the central point: you must have choice - or else it's not a freedom at all. I highly doubt FSF uses word "freedom" as synonym to "improper familiarity"... The AGPL and GPL is exactly the same in this regard, no matter which funny ways you decide to read the four freedoms. How come? Where is fault in my logic? I can show step-by-step that GPL is free software license while AGPL is not free software license. Here is how (just two steps): Note: even the latest edition of RMS's definition explicitly says certain kinds of rules about the manner of distributing free software are acceptable, when they don't conflict with the central freedoms thus reinforcing my idea that freedom 3 is the least important freedom, important freedom, but it's freedom none-the-less, not duty or obligation. And I fail to see how onerous AGPL requirements help with freedoms 0, 1, and 2. It flat out robs me of freedom 0 - and this the most important freedom (that's why when it was added it was added as freedom 0, not freedom 4). I'm sure RMS will change definition soon to make it compatible with AGPL (or it'll just imply that onerous AGPL requirements don't in practice affect freedom 0 at all when they clearly do have impact - and pretty sizable one), but not all people will accept the new definition. Note that DFSG does not have such a requirement (if you can only distribute the program but have no right to run it at all - it's DFSG-compatible), so AGPL is acceptable in Debian.
Care to point any other definition? Posted Jul 27, 2009 18:14 UTC (Mon) by foom (subscriber, #14868) [Link] How come? Where is fault in my logic? [etc]While I certainly have mixed feelings about AGPL, you have not made a very persuasive argument. AGPL's justification for requiring that you distribute source is clearly the same as GPL's: "The freedom to study how the program works, and change it to make it do what you wish." AGPL merely requires that you enable more people to have that freedom than the GPL does.
Care to point any other definition? Posted Jul 28, 2009 6:13 UTC (Tue) by khim (subscriber, #9252) [Link] AGPL's justification for requiring that you distribute source is clearly the same as GPL's: "The freedom to study how the program works, and change it to make it do what you wish." That's merely freedom number 2, it can not trump freedom number 0. AGPL merely requires that you enable more people to have that freedom than the GPL does. AGPL severely limits my freedom number 0 for the sake of freedom number 2. It also imposes "freedom" number 3 effectively making it obligation for no good reason at all. It's like three laws of robotics: everyone agree it's good (even if impractical) laws, but put law number 3 above law number 2 (let alone law number 1) - and you have unmitigated disaster (as Asimov showed in his works). Situation with free software definition is the same: minor changes (like GPL => AGPL change) can convert acceptable thing to abomintaion.
Care to point any other definition? Posted Jul 27, 2009 18:17 UTC (Mon) by hppnq (guest, #14462) [Link] The purpose of all Free Software licenses is to provide freedom to the user, not the developer. The developer, of course, has the freedom to choose which project gets to benefit from her talent and resources. One cannot fail to understand this and nonetheless be more logically correct than RMS, I'm afraid.
Care to point any other definition? Posted Jul 28, 2009 6:20 UTC (Tue) by khim (subscriber, #9252) [Link] The purpose of all Free Software licenses is to provide freedom to the user, not the developer. There are no distinguished users and developers in this world. Everyone uses the software and a lot of people change the software - especially if it's simple set of web scripts. The stated goals are "freedom to run, copy, distribute, study, change and improve the software". But what about the situation where these freedoms are in conflict? RMS showed how copyright is about conflicting goals of the same people - and how big corporations perverted these goals. The same thing starts to happen with free software with the advent of AGPL.
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 8:20 UTC (Wed) by kragil (subscriber, #34373) [Link]
That is your choice, but you don't really have track changes. You just need to make the current running version available.
Just have a script pack the source and put on the web every time you change the production source. It is not rocket science.
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 13:38 UTC (Wed) by tzafrir (subscriber, #11501) [Link]
I hope that this approach will still owrk in the presence of caching.
It's not a conspiracy - it's just ironic twist of fate... Posted Jul 22, 2009 22:04 UTC (Wed) by Wol (guest, #4433) [Link]
That's irrelevant.
If you make the source available as required by the licence, then that's all you're obliged to do.
If the user can't get at the source because of caching, then legally that's called "frustration", where neither side can be held liable for failure.
It's a bit like a contract says "A will deliver the goods to B via the wotsit bridge". If A fails to deliver the goods because the wotsit bridge fell down overnight, B has no comeback against A because A is willing to fulfil the contract - he just can't because the bridge is no longer there.
Cheres,
Launchpad source released Posted Jul 21, 2009 21:21 UTC (Tue) by dlang (✭ supporter ✭, #313) [Link]
all they would have to do is to update the published tarball every time they deploy a new version of code to production.
there is no need to do anything like a tar at the time of the request. for anything other than an interpreted language, a tar wouldn't guarantee that you got the right version anyway.
Launchpad source released Posted Jul 25, 2009 14:32 UTC (Sat) by mdz@debian.org (subscriber, #14112) [Link]
Launchpad solves that particular problem quite elegantly, because Launchpad itself is hosted on launchpad.net. It provides its own source code for download already as a matter of course, because its developers are using launchpad.net.
Link to source Posted Jul 21, 2009 21:11 UTC (Tue) by rfunk (subscriber, #4054) [Link]
Anyone can add a link to their site pointing to whatever code they want, and
claim that they're fulfilling the appropriate license or request or whatever. That doesn't mean that the link *actually* points to the running code.
Link to source Posted Jul 21, 2009 23:42 UTC (Tue) by davi (guest, #18853) [Link]
If you know the application, it is some times easy to detect a new modification.
If such modification is not in the public source code you have caught a license violation.
Launchpad source released Posted Jul 22, 2009 11:38 UTC (Wed) by nix (subscriber, #2304) [Link]
Good point. The source code to the LWN editorial staff should obviously also be released.
Launchpad source released Posted Jul 23, 2009 19:46 UTC (Thu) by SiliconSlick (subscriber, #39955) [Link]
"Good point. The source code to the LWN editorial staff should obviously also be released."
Indeed... complete with the revision history that provides the date for our grumpy editor's initial release.
|
|||||||||||||