Conversation

Notices

1. Embed this notice
   Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Wednesday, 07-Jan-2026 21:51:10 JST Alexandre Oliva

   • Kat Marchán 🐈
   you'll be surprised that, because copyright doesn't reserve the right to run programs exclusively to the copyright holder, copyright licensing (= granting permissions) cannot withhold or delimit that right. there are some clueless license writers fooling themselves and others purporting to have copyleft-like licenses that do this, but all this shows is that they understand neither copyright nor copyleft
   
   the other fundamental problem is that the entire notion of software freedom is that the developers/authors/copyright holders should not have power over users. controlling others through the software they use, even if for a good cause, amounts to unjust power, and such power shouldn't be available. (it also divides the community on what should be ruled out, which is a great way to weaken a social movement)
   
   now, I understand and share the wish to promote socially just causes, and to stop injustice. the way to achieve justice is not to take it on your own hands, imposing your wish onto others by whatever means you can manage. that leads to absolute rule by the most powerful, and I'm sure you're not going to enjoy that. the way to promote social change and achieve justice is by democratically instituted and enforced prohibitions of antisocial unjust behaviors
   
   • 翠星石 likes this.
   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Thursday, 08-Jan-2026 19:51:03 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     are you by any chance conflating the remote users, whose computing the software does, with the operators of the computing service, who run the software for the remote users?
     also see https://www.lx.oliva.nom.br/blog/draft/avoiding-agpl-abuse
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Thursday, 08-Jan-2026 19:51:05 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat I don't entirely agree. The AGPL clearly imposes requirements on users - even if I'm merely deploying AGPLed software I am faced with obligations. We're clearly ok with *some* power over users, as long as the perceived benefits to software freedom are sufficient.

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Thursday, 08-Jan-2026 23:56:26 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     it does apply, and for the scenario at hand, the user/operator already has access to the program's source code, so no further action is required. now, you might want to argue that adding the local git repo and git commit id or tag or whatever to the running software is required (you may be right about that), but that's an absolutely trivial non-onerous requirement, and it's also good software practice. do you really want to argue that this imposes upstream developers' values onto the poor oppressed user? I didn't think so.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Thursday, 08-Jan-2026 23:56:28 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat 13 doesn't have a carve out for the operator and the user being the same person, it still applies

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Thursday, 08-Jan-2026 23:56:30 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     how do the requirements affect someone running the software only for one's own use?!? nobody else can interact with the program, and the user who maintains it for one's own use already has access to the source code.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Thursday, 08-Jan-2026 23:56:32 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat no, the restrictions apply equally to someone running the software only for their own use, so the distinction isn't relevant here

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 00:18:33 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     I stand by my claim. as long as the user keeps the system only for one's own private use, no one else will even know what's going on there, and so the user gets away by default in this de minimis "violation", therefore, there's effectively no developer power over the user.
     
     another way to think of it is to realize that the requirement doesn't apply to the user as such, but to the user as operator. there aren't fundamental ethical objections to imposing requirements onto intermediaries, including operators, to ensure users aren't defined the freedoms they deserve, inasmuchas the requirements refrain from granting the intermediaries power over users, rather than taking their (intermediaries and users alike) freedom away
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 00:18:34 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat I think it's an entirely justifiable requirement, but your original claim was "developers/authors/copyright holders should not have power over users". This is an example of an FSF license imposing the author's will over a user. I'm not saying it's bad, just that your assertion doesn't align with a license we'd both agree is free software.

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 00:49:44 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     you don't get to ignore it, but you get to comply with it by default in the given scenario. or, heck, make it say "get the sources from the local repo, you know where it is" and you're done. there's no oppression there. there's not even an opportunity for an exercise of power.
     
     now, I know you have an authoritarian stance in which you justify imposing your positions on others by such arguments as the ends justify the means, or so it seems to me. I don't see that in the FSF, or in the Free Software movement. but I can see why you project this stance on it.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 00:49:46 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat it's a requirement that applies to me even if I run the software for my own personal use, there's no magical provision under copyright law that lets me ignore it. I don't know why you're arguing with this - it's very clearly a case where the FSF believes imposing something on a user is overall beneficial to free software. I agree with them.

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 01:27:08 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     where is that quote from? you wouldn't (and shouldn't) ever find it in a licensed aligned with free software philosophy.
     
     as for whether it's authoritarian, I'm not surprised we disagree on that point.
     
     CC: @zkat@toot.cat
     
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 01:27:10 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat I don't know how "You must abide by this requirement to use the software" is anything other than a (mild!) exercise of power over the user. If I fail to adhere to that requirement I'm breaking the law. I don't think it's authoritarian, just as I don't think people advocating for things like ethical software are authoritarian - as communities we've simply made different decisions about what kinds of impositions are acceptable.

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 03:57:00 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     you access the server and use it.
     
     or, if you want to run your own server, you install the software there and use it.
     
     oh, wait, you said without abiding by the requirements of section 13. both of the above abide by them.
     
     so you want to modify the program? fine! if you insist, adjust its message to tell yourself where you put the sources and you're done.
     
     just don't make such a fuss about it. with the time you've put into this teapot tempest, you could have written the message for yourself a hundred times.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 03:57:02 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat it's a paraphrase of the AGPL - how do I use AGPLed software without abiding by the requirements of section 13?

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 04:02:06 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     your paraphrase sucks BTW. it's a wild misrepresentation.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 04:23:40 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     you're still missing the point that you abide by it by default, unless you choose to modify the program. if all you want to do is use it, you don't have to go out of your way at all. just like GPLv2's and GPLv3's wordings that you don't have to accept the license. you need to abide by it to modify, copy, distribute, publish, that are activities reserved by copyright law. GPLv3 and AGPLv3 even affirm explicitly the unlimited right to run the unmodified program. that sentence should not be necessary for you to have that right as a user, but since so many people are confused by propaganda by copyright maximalists, it doesn't hurt to have it there.
     
     CC: @zkat@toot.cat
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 04:23:41 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈

     @lxo @zkat right, it's a requirement that I have to abide by to use the software. It's not onerous, but it's unambiguously a requirement and it's unambiguously a free license regardless.

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 06:35:23 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     if you removed the download notice from the program, it's no longer unmodified.
     
     CC: @sally@freesoftwareextremist.com
     
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 06:35:25 JST Matthew Garrett

     in reply to

     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     @sally @lxo you're saying that deploying unmodified AGPLed software in a way that removes the download notice is license compliant?

   • Embed this notice
     𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙 (sally@freesoftwareextremist.com)'s status on Friday, 09-Jan-2026 06:35:26 JST 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     @mjg59 @lxo @zkat
     
     No, this is a load of bullshit and Alexandre told you that already. You don't have to do shit if you're the only person using the software, because nobody other than you can enforce its terms, copyleft licenses do not matter unless you're publicly distributing software, personal use is not such thing.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 06:35:27 JST Matthew Garrett

     in reply to

     • Kat Marchán 🐈
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     @sally @lxo @zkat what? I like AGPL, it's just a license that has (minor! Justifiable!) requirements that have to be followed even if you're the person using the software

   • Embed this notice
     𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙 (sally@freesoftwareextremist.com)'s status on Friday, 09-Jan-2026 06:35:28 JST 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     @mjg59 @lxo @zkat
     
     You're fucking dumb or just trying to die in a hill hating AGPL no matter how irrational it is, and Alexandre should've just told you that rather than waste time explaining this, which is something even the average Joe would understand.
   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 07:23:09 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     I see you've retreated from the claim we were discussing before. Good. My job here is done. I won't speculate on how a court would rule on such an attempted circumvention in a (public?) performance.
     
     CC: @sally@freesoftwareextremist.com
     
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 07:23:11 JST Matthew Garrett

     in reply to

     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     @lxo @sally the software is unmodified, I'm placing a proxy in front of it that modifies the HTML

   • Embed this notice
     Alexandre Oliva (lxo@snac.lx.oliva.nom.br)'s status on Friday, 09-Jan-2026 08:04:13 JST Alexandre Oliva

     in reply to

     • Matthew Garrett
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     • robryk
     the requirements you so insistently hinged on only apply when you modify the software.
     
     now you've moved to a different scenario. it's unclear whether this new scenario still involves a single operator=user, or whether the software is being publicly performed for others to use.
     
     in the former case, I stand by my earlier claim: nobody (aside from you, it seems) cares if you run the program however you like, you've already got the corresponding sources before you even install the program in whatever weird scenario you can concoct.
     
     now, if you're running the software for remote users, then they deserve freedom, and AGPL aims at making sure they can get the freedom they deserve, while you are not the user, you're an intermediary trying to plot ways to take their freedoms away, and that's not a freedom, it's power over the remote users, and nobody deserves that kind of power, so it's fair game to curtail it.
     
     CC: @sally@freesoftwareextremist.com @robryk@social.wuatek.is
     
     翠星石 likes this.
   • Embed this notice
     Matthew Garrett (mjg59@nondeterministic.computer)'s status on Friday, 09-Jan-2026 08:04:15 JST Matthew Garrett

     in reply to

     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     • robryk

     @sally @lxo @robryk If I can't put a proxy that performs arbitrary modification of output in front of unmodified software, that sounds like you're controlling the way I use it (in a very insignificant way, to be clear!)

   • Embed this notice
     𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙 (sally@freesoftwareextremist.com)'s status on Friday, 09-Jan-2026 08:04:16 JST 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     in reply to

     • Matthew Garrett
     • robryk
     @robryk @mjg59 @lxo
     
     The intent of proxying bits of software is to de-facto modify or deliver a modified version of it through the network, if you're trying to cheat around to avoid having to comply with the terms of the license you'll need good luck in court to convince attorneys and judges that you didn't try to do something that the license deliberately told you not to do.
     
     This practice could be called "going against the spirit of the license", because that's exactly what you're trying to achieve, TiVO tried similar shenanigans and everyone with more than one digit IQ immediately understood what they were trying to do and hated it, also the license got patched right after.
     翠星石 likes this.
   • Embed this notice
     robryk (robryk@social.wuatek.is)'s status on Friday, 09-Jan-2026 08:04:18 JST robryk

     in reply to

     • Matthew Garrett
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     @sally @mjg59 @lxo Would you mind pointing out where it has them covered? From reading it years ago it seemed ~equivalent in that regard to GPL, and from reading GPL it very clearly stated that it applies only to distribution and not to usage.
   • Embed this notice
     𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙 (sally@freesoftwareextremist.com)'s status on Friday, 09-Jan-2026 08:04:19 JST 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙

     in reply to

     • Matthew Garrett
     • robryk
     @robryk @mjg59 @lxo
     
     > If that also applies to AGPL, then placing such a proxy is totally fine.
     
     Good luck convincing everyone that it wasn't done maliciously, a scenarios which AGPLv3 has covered, otherwise sure, you can.
     翠星石 likes this.
   • Embed this notice
     robryk (robryk@social.wuatek.is)'s status on Friday, 09-Jan-2026 08:04:20 JST robryk

     in reply to

     • Matthew Garrett
     • 𝅙𝅙𝅙𝅙𝅙𝅙𝅙𝅙
     @mjg59 @lxo @sally I think that the GNU position is that one doesn't need to agree to GPL to use the software (see e.g. https://flameeyes.blog/2020/03/23/the-gpl-is-not-an-eula/ for a better explanation than I can offer). If that also applies to AGPL, then placing such a proxy is totally fine.
   • Embed this notice
     翠星石 (suiseiseki@freesoftwareextremist.com)'s status on Friday, 09-Jan-2026 21:52:20 JST 翠星石

     in reply to

     • Matthew Garrett
     • Kat Marchán 🐈
     @mjg59 @lxo @zkat Why are you claiming things the AGPLv3 does not require?
     
     Section 13 only applies "if you modify the Program" and merely using the software is not modification.
     
     Any even if you do modify it; "your modified version must prominently offer all users interacting with it remotely through a computer network" - clearly there are no users - there is only you, the host - so you don't need to do anything.
     
     Besides, "(if your version supports such interaction) an opportunity to receive the Corresponding Source of your version" - your modified version doesn't support providing yourself the source code, so you don't need to do anything regardless.
     
     Regardless, you cannot modify the program without ensuring that you end up with the source code in the end, so section 13 is a total no-op when it comes to private usage.