DOC Clarify right to contribute and paid-platform disclosure - #34490

@DeaMariaLeon, would you mind having a closer look?

Two thoughts on this:

1. it is annoying to be doing this kind of unpaid labour, in particular without knowing it
2. our contributing docs and PR template are getting ever longer and more overwhelming to read

I'm not sure what to do about 1, beyond what we already do: close PRs and issues that seem to be low effort and not solving a problem we care about (at the moment). I'm wondering if we need a special case for this kind of low effort and priority work, or if we need to get a bit more organised internally to make sure everyone knows what is "low priority" and feels confident to close those contributions with little mental overhead.

My assumption is that abusive people will either not read our contributing docs or simply tick "the right boxes" to allow them to proceed. By having big text walls and "legal" questions we then put off the conscientious contributors. For example, having to fill out a CLA makes it quite a bit harder to make a contribution: I now have to explicitly deal with this topic, it contains legal text, I probably ave to involve my employer (and their lawyers) who often have never seen a CLA before in their life, etc. At best this means weeks pass before someone can contribute, at worst the lawyers do what lawyers do: if in doubt say no.

I think we should be fine, if we make a reasonable effort to inform people about the licensing requirement, individual contributors make a reasonable effort to inform themselves/figure out their legal situation and we are responsive when someone later comes to point out that something went wrong.

TL;DR: I like the information about the licensing in the various contributor guide, FAQ, etc documents. I would remove the FAQ and paragraph text specific to "are you doing this for a bounty?".

Related to my discussion about https://github.com/openclaw/clawsweeper with a friend, he pointed out that they use a bot to ask contributors to point to evidence that this is a useful/welcome contribution. For example, they can link to existing issue(s) where maintainers have said "yes do this" or reddit threads or such to proove that there is a "real reason" for this contribution. I think this is a cool idea, maybe not using a bot for it, but the "please include material to convince us". Maybe going in this direction will also reduce the number of "bounty" contributions that are not also useful for the project. Concretely: ask contributors to include proof that this is a useful contribution.

Having thought about this a bit more, I think I am against making contributors sign a CLA/tick a box/variation of this.

I'm not sure it will help reduce the "contribution for money" problem and it will add a hurdle for good contributors.

On a more philosophical note, sustainable open-source is about exchanging value between maintainers and contributors. There has to be "something in it" for both sides for the project to continue for a long time. If a contributor can get paid for their work that is great, it is the simplest way to benefit from the contribution. You can exchange the money for basically anything else you might need. Things like fame&recognition which is often what contributors gain are much harder to exchange for other things.

What is missing from the "solve problems for AI training" problem we are discussing/reacting to here is the value for the project. The ideal contribution would be something where the contributor gets paid and the project benefits as well. If someone then also trains a LLM on this doesn't bother me. We are all benefitting/the rising tide lifts all the boats. So I am thinking that we should work more on how to evaluate the "benefit for the project" part. Asking contributors to show why their contribution is useful, having a clearer way to decide as a maintainer, etc. I think this is a harder task, but the thing we should focus on.

Having thought about this a bit more, I think I am against making contributors sign a CLA/tick a box/variation of this.

I've only read this after answering to Gael's suggestion here: #34489 (comment) earlier today.
I will treat your reply as a maintainer veto and close this PR if you don't thing anything is worth keeping. Please let me know in case you're okay with keeping anything from this PR, in particular this:

Do not open pull requests mainly to meet the requirements of a third-party paid platform
(for example a bounty or paid "quest", or similar programs that compensate you for
opening PRs or producing training data). Do not use scikit-learn's review process as
unpaid QA for an external company.

The ideal contribution would be something where the contributor gets paid and the project benefits as well. If someone then also trains a LLM on this doesn't bother me. We are all benefitting/the rising tide lifts all the boats.

I don't believe in the "the tide lifts all the boats" narrative. The expense open source projects pay is dedicating too much attention to contributions whose economic interests remain hidden. If it were all open and clearly communicated, my assessment would be very different.

I would keep the entry in the FAQ and the later part in the contributing guide. My only thought there is if we could find a different formulation than "right to contribute".

Great, I will merge it then (since we have two approvals).