5 ms·
Seems like you still don't understand - the license designed to scare people like you :) Mostly big corps.
by paulcarroty 6y ago
Seems like you still don't understand - the license designed to scare people like you :)
Mostly big corps.
- knorker 6y agoIt's designed to scare away open source hobby programmers? And users? This seems like "avoid success at all costs". The way I see it there are two kinds of potential AGPL software users and contributors: 1. Those who have not read and/or understood the license, and just go with it. They likely will be fine, because who's going to sue John Doe who built his first website in PHP, but didn't publish his AWS script for launching mongoDB? 2. People who nope the fuck out when they read the license, and don't become contributors or users. (1) are violating the license. (2) are not using it at all. So what an AGPL software author does is take the pool of potential users and contributors, and divides them into those who nope out, and those who violate the license. And is that really good honest work you can be proud of? That reminds me more of the music industry of 15 years ago. They gave their customers two options: Have to deal with spinning plastic, or pirate. In this comparison the former is "nope, guess I won't listen to that band, then". AGPL software authors often make the point you are: Well nobody is using plastic, and so what if they pirate the software? They're not profiteering on it, and it was free anyway. Just make a license that says it's not allowed to be used by companies of size greater than X.
- commoner 6y ago> Just make a license that says it's not allowed to be used by companies of size greater than X. That kind of license would not be free or open source. Such a restriction is not allowed in both the Free Software Definition and the Open Source Definition. > A free program must offer the four freedoms to any user that obtains a copy of the software, provided the user has complied thus far with the conditions of the free license covering the software. Putting some of the freedoms off limits to some users, or requiring that users pay, in money or in kind, to exercise them, is tantamount to not granting the freedoms in question, and thus renders the program nonfree. https://www.gnu.org/philosophy/free-sw.html.en https://www.gnu.org/philosophy/free-sw.html.en > 5. No Discrimination Against Persons or Groups > The license must not discriminate against any person or group of persons. https://opensource.org/osd https://opensource.org/osd The AGPL is both a free software license and an open source software license because it ensures certain freedoms for end users, including source code access. Intermediaries who want to host modified AGPL-licensed software, but do not want to release the source code for their modifications, have the option of negotiating a commercial license with the original developer. The commercial distribution of the software would be nonfree, but it would fund development of the free software.
- knorker 6y ago> That kind of license would not be free or open source. I agree. And I'm saying AGPL is de facto not free software, because ~100% of serious (even hobby) installations violate it, even if it's free software de jure. And independent of this I also think AGPL violates freedom 0. E.g. the AGPL explicitly forbids you from running a service combined with proprietary software. You are NOT allowed to run a service by combining AGPL software with a binary blob you don't have source code to. How could this possibly be "The freedom to run the program as you wish, for any purpose"? If I'm not allowed to run X with Y, how is it "as you wish, for any purpose"? > have the option of negotiating a commercial license with the original developer. Only if ALL contributors (aka copyright holders) agree to this. Copyright reassignment is needed from the beginning of the project, for all contributions.
- 2Gkashmiri 6y agoI am in no way an expert, just a novice in gpl licensing but I have some questions. Why do you say agpl is not for free software / hobby ? From my understanding which could be wrong, is that you are free to use "this" software if you give the same freedom to your downstream work and any modification you do must be given back to original work. Why is that difficult ? If I am using libreoffice and I develop a fix for bug that I found in my usage, why not give it back to the entire community so that the software is made better for the entire mankind? Second, which is more important than anything hobbyist use, if I am a for profit corporation looking to earn money, why should I be allowed to take existing code built by God knows how many volunteers, make slight changes and market it as "my" product and totally ignore the work done by these people? From what I have been reading over time, no one is going to come after a "serious hobbyist" if they ran GPL code in their projects having proprietary blobs. That is just scare mongering. Leeching off of work of community and passing it as own work without passing modifications and improvements to community is what the issue is and unless you want to do that, fine. Use commercial or software that is permissive only.
- knorker 6y ago> Why do you say agpl is not for free software / hobby ? I mean it's not free software, even in the context of the user using it as a hobby thing. Is that what you're asking? > From my understanding which could be wrong, is that you are free to use "this" software if[…] So that's a problem right there. You are free to use this software "if…". FSF's free software definition includes freedom 0: "The freedom to run the program as you wish, for any purpose (freedom 0)" So I guess I can flip the question around to you. How do you mesh these things: 0. The freedom to run the program as you wish, for any purpose 1. I wish to run this software intertwined with some proprietary code 2. The purpose is to run it as part of a service offered to the public 3. The software is AGPL 4. The AGPL states that I cannot combine 1-3. 5. But I wish to combine 1-3 together, and 0 says I can run it as I wish for any purpose. > Why is that difficult? It's not difficult, it's logically impossible. My conclusion is that AGPL throws away freedom zero. And if that's out, then the license could just say "may not be used by companies above size X". And it's not free software. > If I am using libreoffice and I develop a fix for bug that I found in my usage, why not give it back to the entire community so that the software is made better for the entire mankind? Indeed. The problem here is that what if it wasn't a bugfix, but a feature where you added support for your company's SSO system, and used libraries that are not open source? Or what if you're writing a script to launch openoffice in your VMware cluster, to convert from .doc to .docx. You now have to publish that script, if in any way connected to any service that anyone else would possibly and indirectly access. Like, you can't use openoffice as a report generator in this way without publishing the script that launches openoffice on your cluster system. And if you edit /etc/init.d/mongodb, you do have to publish that new version. > Second, which is more important than anything hobbyist use, if I am a for profit corporation looking to earn money, why should I be allowed to take existing code built by God knows how many volunteers, make slight changes and market it as "my" product and totally ignore the work done by these people? I don't think your scenario here is right. All open source licenses require copyright to be clear. Even BSD license without advertising clause doesn't allow you to claim it as your own. But reading your question as only applying to services. Hmm… how is this different from you running a blog and not acknowledging all the work that went into the Apache webserver? Why does the public have the right to your startup scripts and apache.conf? Why do they have the right to your cronjob script that installs a new letsencrypt and restarts the webserver? Because this is what AGPL demands. Nobody thinks Amazon invented xen or Linux. Nobody operates in a vacuum. Even Nobel prize winners didn't derive physics from first principles. The AGPL is the most extreme. I don't think you should consider it the norm. There are many popular licenses that deliberately chose to allow modified binary releases. BSD, for one. And that's fine. > From what I have been reading over time, no one is going to come after a "serious hobbyist" if they ran GPL code in their projects having proprietary blobs. Don't confuse GPL and AGPL. Doing that is entirely allowed by the GPL, and thus there's nothing to "come after". Because GPL has freedom 0. AGPL though, yes you're right nobody is going to go after you and your blog. But you're still in violation. You are pirating that software, because if you don't follow the license then you have broken the contract under which you're allowed to have it, and it defaults to you not. So you're not actually using AGPL. You're just pirating. But no, nobody's going to come after you. Unless they're a dick. But you're in the wrong, so if you do get sued you'll likely lose. > Leeching off of work of community and passing it as own work without passing modifications and improvements to community is what the issue is and unless you want to do that, fine. I do recognize the problem. The best weapon against it that we have is that it's expensive to maintain your own fork, and to keep merging it with upstream. Part of the reason Google et al contribute to the Linux kernel so much is the selfish reason that it's MORE expensive for them to maintain their own fork with local patches they have to apply every time there's an upstream release. It's why many companies contribute. I've done it too, and small and big companies. It's just cheaper to upstream your patch, and have it show up a while later in an apt-get upgrade. That won't always work, though. But choosing AGPL all but guarantees no corporate users or contributors. And many many non-corp too. And your project is all the worse for it. And do you want to make good software for the world to use, or do you just want to stick it to the man? To end on a similar note to you: if your goal is not to make something good, but to stick it to the man, then fine, use AGPL.
- luord 6y ago> It's designed to scare away open source hobby programmers? And users? Just giving my two cents: I don't have any trouble contributing to AGPL projects. Mind you, I haven't contributed much in general, but still. As for using AGPL projects: Well, I wouldn't want to offer someone else's project as a service and poach potential users for the maintainers in the first place. But if I did, I wouldn't mind contributing back any modification I made. Now, using them as dependencies—for just one part of an application that you want to keep otherwise private—is a little trickier because I'm neither a lawyer nor has the AGPL been tested in court as far as I know. The crux of the issue lies on whether dynamic linking makes a piece of software derivative; the FSF thinks it does (at least from what I read in their FAQ), but that also hasn't been tested in court (again, AFAIK) and one's interpretation could be entirely different. If it isn't derivative, then there's no problem with using AGPL software strictly as dependency. And, if it is, well, in the world of microservices you could just create a thin wrapper over the AGPL library, run it as a service, and release only that. Then again, if one assumed that linking doesn't make a work derivative, then there's not much difference between the GPL and the LGPL (once again, AFAIK). It's a nuanced issue, to say the least.
- SOLAR_FIELDS 6y agoIt’s rather effective in that regard. When I worked in BigCorp, the open source contributions had to be reviewed. There was a blanket policy that anything GPL/AGPL was not to be touched under any circumstances.