Many people find the AGPL to be a nonfree license, myself included.
I don't understand this stance.
From the point of view of the user, the AGPL offers the same freedoms as the regular GPL. If you use the program, you have the right to see and edit its source code, regardless if the program runs on your own computer or on another computer.
How can anybody call this feature "nonfree" is beyond me.
Granted, the license only asks you to publicize your modifications, but in software that's basically your "first-born".
If you're just talking about the freedom to run things unmodified, basically every software you can download online without paying is "free", and obviously AGPL fits that bill for the uncapitalized "free" software. But that's not what we're talking about here, right?
The situation is funny that AGPL is widely derided as both being to weak and being to strong.
If that is the case, I don't see how the AGPL qualitatively infringes on any "freedoms" that the GPL doesn't already restrict: you must keep derivatives free. Of course one may argue that the GPL is nonfree when compared to BSD/MIT style licences, in which case fair enough.
I think the argument of FOSS people who don't like it would be something like, as a user, just running the software on my machine can potentially create a legal obligation to set up source distribution, if, like, I forgot to block a port.
In any case, does the AGPL requirement to provide sources apply even if you are running an unmodified version? In other words, would merely pointing to the original repository not be enough?
Regarding the risk of accidental legal obligations, I suppose it depends on the details and technical wording of the licence.
If the requirement broadly covers any use of the software on the machine (e.g. some backend service that happens to help your public webserver to stay online) I can see the discomfort, especially if one must explicitly provide or link to unmodified sources too.
On the other hand though, if the requirement is limited to actual provision of the software's functionality to 3rd parties, I would argue that if someone accidentally provides it due to forgetting to block a port then they have much bigger problems than the AGPL.
https://www.gnu.org/licenses/agpl-3.0.en.html
It just says interact remotely through a network. Interpreting what this means precisely is left as an exercise for the reader / the reader's lawyer.
forgot to block the right port
if you get notified of a security breach caused by your own incopetence, you secure it. which would effectively be the highest harm apgl can inflict on you already: stop hosting the service. which was your point all along.
for a popular program that's not very likely, but lost source code is the bane of software development, so hacing your own version available would be better, at least as a backup
distributions keep archives. Just use a distribution and you're fine.
Not true: only if you made local modifications.
> I forgot to block a port
Also not true. You have plenty of time to close that port.
"""your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice."""
If you expose services on the Internet by mistake you are going to have 1000x bigger problems than the risk of being sued successfully for a honest mistake.
That said, FOSS authors normally aren't the kind of people that pushes to those crazy maximalist interpretations of documents. So if any risk exists, it's not large.
It's not entirely clear what counts as derivative work with regards to GPL. There is a lot of corner cases and workarounds in the real world. It becomes even less clear when you widen the definition of distribution.
As a corporation, I might not want to use it in conjunction with my code, but that's the point. You are free to do what you want, but you in turn must share your code.
Or you engage with the owner to make a financial arrangement for a separate license.
This is like nonfree usage restrictions on otherwise free software, like the famous (if unenforceable) “can not be used for evil” license clause.
therefore the AGPL shouldn't force you to share code if the program runs on a server. but it should force you to share code if the program can be accessed by a user.
this may even apply to running something like a kiosk device where users interact with a program on a device they don't own.
I.e. the gpl prohibits you not sharing code if you distribute the binary. What is the philosophical underpinning of "sharing binary" not being a restriction but "ethernet" being a restriction.
The evil thing is not similar because you are discriminating against a field of endevour (similar to licenses that say educational use only, non-commercial only or non-military only). It seems to be a very different type of restriction.
That's not even remotely near what it is.
I've worked for several large corporations and were in the throes of a major project. Using an Open Source tool would've cut a lot of our dev times by many hours. Three out of three times, all three team leads and managers said "no" to using open source. We had to build several large components from scratch, which did lead to extended development timelines, a ton of working late at night and several "crunch" weekends.
Their rationale? They didn't want to have to maintain a piece of software that was open source. The resounding opinion was it was "dangerous" and not safe - because someone could check in a malicious piece of code and that could end up in our very large, enterprise application when we needed to update it. Nobody wanted to be on the hook if that ever happened.
In the end, it seemed like it was more about upper management types playing CYA instead of trying to put out a solid product.
Do you have evidence?