While your assertion is true, it's not the only reason (or even the main reason) for CLAs in OSS.
IANAL, but this has actually happened in the past (the big case I'm aware of predates open source but not free software -- google for Diku MUD). Let's say that someone infringes on the license. So you take them to court to get them to follow the license. If you have a CLA you are the copyright holder so you can just do this. If you don't, to make your case you have to round up virtually everyone who has ever contributed to your project because they are copyright holders too. So, in effect, it doesn't matter if they infringe on the license or not because you can't hold them to it.
How so? As long as your code is also being used, you can sue for copyright infringement on your work. I don't see why you need to call up anyone else.
I'm not a lawyer and can't explain why things are the way they are. I can only inform you that this is the way things have shaked out in the past.
A prime example is the SFC vs. VMWare lawsuit, where SFC represents a single contributor to Linux Kernel (Christoph Hellwig).
No, if you hold a valid sublicensable license (which includes any F/OSS license, as that's rather the point of Free Software/Open Source Software), and make a derivative work as authorized by that license, you hold the copyright on the derivative work and can enforce it without any participation by the upstream copyright holders.
You may run into difficulty if the offender can show by a preponderance of the evidence that their allegedly-infringing work is derived from the upstream rather than your derivative work, in which case the upstream would have to take action.
This is demonstrably untrue - all the Linux kernel GPL enforcement cases have been brought by individuals without involving any other kernel copyright holders.
On the other hand, without signing over your rights, you can effectively hold up a project from moving forward. For example, if a project wanted to move from GPLv2 to GPLv3, IIRC, you can't do that without the permission of all contributors unless they have signed over their rights. And as @trestletech has pointed out not all contributors are created equal[1].
One the one hand, you have people making significant contributions (and you don't want to just stomp all over them), on the other hand you could have people making tiny contributions, yet are holding up everyone else.
[1]: https://github.com/datastax/nodejs-driver/commit/6f3c6b2f454...
Obviously the devil's in the details of the terms of the CLA. I had previously thought the purpose of the CLA was to allow the projects organizers to make copyright claims in court, not to allow them to re-license.
http://www.cnet.com/news/linux-to-gplv3-a-practical-matter-n...
As long as contributors are licensing their contributions under the terms of the GPLv2, they're also licensing them under all future versions.
No it doesn't. The relevant part of the license is:
Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.
ie, it's optional to use the "or any later version" language.
You only need them to license their code under "GPLv2 or later" while you license the project only under "GPLv2". This way you're free to upgrade the license if and when you want.
The main exception I can think of is Debian. I'm not sure what the history is there or why their lawyers came to a different conclusion than everyone else.
(Disclaimer: I am the DataStax founder but I am not involved in the drivers team or their CLA.)
Nonsense. Red Hat don't require a CLA for any of their projects. The Linux kernel doesn't have a CLA. Libreoffice doesn't have a CLA. Mozilla don't require a CLA. The Python CLA might have that name but is actually just an explicit statement that the contribution is licensed under certain terms, which makes it entirely unlike the Apache or Canonical ones.
"I'm not sure what the history is there or why their lawyers came to a different conclusion than everyone else."
Most projects with CLAs have them because they see other projects with CLAs and it seems like the safest option. The common justification of "We need to know that the contribution is authorised" is more elegantly solved by the Developer Certificate of Origin approach, and the secondary justification of "What if an original contributor revokes permission to distribute" is only a problem in countries with string moral rights and a CLA isn't going to save you there anyway.
The FSF has copyright assignment for strategic purposes (they want to be able to launch enforcement lawsuits without requiring that the original authors all register their copyrights), and Canonical have a CLA because they want to be able to release proprietary versions of their code without granting anybody else the same freedom. The majority of cases are just entirely unnecessary.
CLAs certainly have pros/cons - it gives the project the right to enforce the license independently, but it also gives the project the opportunity to relicense the project without consent from the contributors. That may be a problem, especially when the project is driven by for-profit organizations.
Linus Torvalds: To be fair, people just like hating on Canonical. The FSF and Apache Foundation CLA's are pretty much equally broken. And they may not be broken because of any relicencing, but because the copyright assignment paperwork ends up basically killing the community. Basically, with a CLA, you don't get the kind of "long tail" that the kernel has of random drive-by patches. And since that's how lots of people try the waters, any CLA at all - changing the license or not - is fundamentally broken.
While, IMO, the FSF & other non-profit agencies are an appropriate holding body, I would seriously have issues with a for-profit company asking for a CLA for open source work. I would instead reply with a request for a Statement of Work & some cash...
Edit: Oops, no it looks like it's with DataStax. Then I can understand why people wouldn't sign it.
The fact that your contributions are now effectively owned by DataStax and then licensed back to the people who actually wrote them(!) gives me a very bad feeling, like you've just become an unpaid employee of DataStax.
I will fully admit that I haven't read DataStax's CLA in full, but that is because I cannot without giving them access to my GitHub account...
PS - I understand that the FSF does this also, but the FSF isn't a for-profit company so apples and oranges in my view.