Don't sign a CLA
@sir I would judge each CLA on its own merits, rather than saying 'no' outright. They can be an overreach, but they can be part of a mature and well-managed project's governance. It depends on what it actually says.
@marxjohnson do you have some examples of good CLAs?
@sir I think the Apache CLA is quite good. It makes it clear what you can expect from the ASF, has an explicit declaration that you can legally license the work to them, and it addresses non-copyright issues like a patent license. It's good due diligence for projects, especially if they might want to provide commercial services around the software.
@marxjohnson just gave it a read, it's better than most CLAs but imo it's still heavy handed. The same problem is solved more simply by, for example, Linux
@sir there are a lot of cases here  and they don't all seem bad? Mozilla for example. If the license has been around forever, the risk of needing to relicense seems minimal, but for newish licenses I think I'd want to leave open the possibility of fixing mistakes.
@skybrian relicensing can be done without a CLA, you just need to contact the original authors, tell them the plan and obtain their consent, then rewrite any code from anyone who doesn't consent. Generally the latter camp is pretty small, if your cause is just. In the vorbis example, the argument is strong for encouraging the adoption of free codecs at the expense of loosening the license of the implementation.
cmpwn.com is a private Mastodon instance for friends of SirCmpwn.