Do we need the CLA for documentation PRs?

[encukou]

No description provided.

[pradyunsg]

Based on the notes in https://docs-community.readthedocs.io/en/latest/monthly-meeting/2022-02.html, the actionable item here is for someone to talk to PSF's general counsel about whether we need to do this.

There's also the aspect of figuring out if this is a less of a documentation-specific concern and whether the broader efforts to improve the CLA process will make this requirement less problematic. :)

[CAM-Gerlach]

☝️ @VanL

There's also the aspect of figuring out if this is a less of a documentation-specific concern

From a non-lawyer perspective, there seem to be two aspects here.

First, there is the question if a contribution is an expression sufficiently creative, original and substantial to be protected by copyright; if not, then a CLA would appear to be mostly redundant. This applies equally to the code and the docs; in either, a single-word typo fix is probably fairly safely non-copywritable, while a large 100-line addition of the contributor's original work very likely is. However, there is no hard and fast rule, so it would require either a licensed attorney judge every one (and even then, it is ultimately up to the courts), or, if they see it fit, legal counsel establish some clear, conservative guideline (that can presumably be unambiguously determined programmatically by the bot).

The other side of things is whether the docs, being stored in the same repository as CPython itself and nominally under the same license, but distributed separately, is worth not being afforded the protection the PSF's specific CLA grants in exchange for potentially being more accessible to new contributors. On the technical side, this would basically involve the CLA bot not triggering on PRs that only touch files of a specified type in the documentation directory, more or less. However, the real question is on the judgement call side, which is a decision for PSF leadership and counsel.

whether the broader efforts to improve the CLA process will make this requirement less problematic. :)

From my IANAL reading, the PSF's CLA seems pretty non-invasive as CLAs go; it mostly reads like a DCO rather than the more common type of CLA that assigns copyright interest to the organization mandating the CLA. As such, it would seem most of the pain would likely come from the complexity and latency of the current bot, rather than the terms being onerous (at least, that was the case for me when I signed). I seem to recall a recent issue with a lot of discussion on this point, though I can't find it now—I believe, however, members of this WG (including @Mariatta , IIRC) were pretty involved there.

[CAM-Gerlach]

See this relevant Discourse thread where @VanL weighed in on a related issue.