So let's be clear: the FSF have been completely negligent in their documentation of the AGPL's effects, and how people should use it.
It is *not* a magic SaaS loophole fix.
Please, *please* do not use it unless *you* understand it better than the FSF do.
Conversation
I've violated the AGPL just by installing a package in my distribution (Gentoo), by some interpretations.
Other distributions violate the AGPL themselves, by the same token.
Or maybe they don't?
Nobody knows. It's a giant mess.
And none of this is tested in court.
1
1
37
So is there some other license that you should use instead?
I don't know. I haven't seen a good one yet.
Unless you *really* care about this issue, very deeply, to the point of being willing to hire legal counsel to analyze your options, I suggest you don't bother.
1
1
16
And if you *do* care about this issue that much, I am very interested in seeing what you come up with, and how you manage to address all these issues.
Please document it properly, like the FSF didn't.
1
1
22
On the other hand, the AGPL is *perfect* if you are a commercial enterprise who wants to get all the Free Software brownie points by releasing source, while also providing SaaS or commercial options and guaranteeing nobody else can. The toxicity of the AGPL is *perfect* for this.
3
7
33
I guarantee no other company will dare touch your software in any "dangerous" way if you do that.
Of course, if you use the AGPL like this, you also need a CLA requiring external contributors to assign copyright to you (or equivalent). Which makes you a leech on the ecosystem.
1
21
Any company requiring external contributors to assign their copyright to them, or full relicensing rights, is effectively using the free software ecosystem as unpaid labor. It's just as bad as the music industry exploiting artists to get them to sell off their rights.
3
3
30
As you may know, the FSF is one such "company", requiring copyright assignment for many of their own projects. Funny, isn't it?
But it goes further than that.
1
1
18
Don't forget they *also* have successfully convinced the majority of the GPL ecosystem to use "or any later version" wording, which gives them perpetual relicensing rights over the *vast majority of GPLed code in existence*.
1
2
22
You think Google's CLA is bad? The FSF have convinced everyone to effectively CLA their code to the FSF and convinced them that this is a *good thing* and it's fine because they can be trusted.
They are not trustworthy.
4
4
34
Replying to
Google's CLA doesn't require copyright assignment. It gives them a permissive license for your code and an explicit patent grant. They give you the same thing by releasing their projects under the Apache 2 license.
The terms would be unfair if they released their code as GPL.
Replying to
Why do they need the CLA if the license already gives them the same thing? You don't need a CLA in order to get rights for a contribution licensed under a given license.
1
2
Replying to
Have you looked at cla.developers.google.com/about/google-i?
Compare it to apache.org/licenses/LICEN, which even has a section on contributions. It's nearly the same.
They require a CLA because they want you to explicitly agree to those kinds of terms and have it on record that you agreed to it.
1
3
Show replies