Le mercredi 18 décembre 2024, 11:07:25 heure normale d’Europe centrale Dave
Cridland a écrit :
> On Tue, 17 Dec 2024 at 22:12, Peter Saint-Andre <stpeter@stpeter.im> wrote:
> I think there's a third option, which is to move the responsibility for
> processing non-GitHub PRs onto the submitter, and have them find a
> volunteer to handle it each time (a Document Shepherd, additional Author,
> or whatever). I worry that if there's a body on the Editorial team who
> handles this very rare case, it'll likely mean that when it's needed, we've
> mostly forgotten who has this role, or they've drifted away, or whatever.
> Once submitted, it's the Author's responsibility to do the updates by PR as
> well, so this is a task for the lifetime of each XEP (like the Author,
> really).
>
> If a submitter cannot find anyone on the standards list or amongst the
> people they know willing to help handle this part of the Author role, then
> I think that suggests quite a bit about how much interest there is in the
> XEP.
>
> Furthermore, I think this is essentially a mild formalisation on what the
> existing de-facto process is; so absent a strong reason to chnage the
> existing process, I'd rather we document something that matches.
>
> (I'll make a PR against XEP-0143 with concrete text on Friday)
I think that making it nearly-mandatory to register an account on a service
from a private company is a problem. It's totally understandable that people
don't want to create an account on GitHub (I've avoided it myself for a long
time), and asking potential first-time contributors to look for help themselves
can be highly discouraging.
Totally agreed.
At the very least, there should be an easy process to ask for help, a list of
people to contact, or an automated message requesting assistance when a patch
is received by email. As you mentioned yourself, this case is currently rare,
so it shouldn't be a huge problem.
Is something like "Do a PR, but if you can't, send it to standards@ and ask for help?" enough here? Currently, Florian is happy to handle all such requests; but I don't think it's fair on him to rely on his continued availability and kindness.
I don't say that accepting patches by email is a must, but there should be an
option that doesn't require creating an account on a private company's
service, even if it's not used often.
I do 100% sympathise. However, options such as running our own version control, or insisting that the Editor handles emailed patches and submissions, seem to put the additional load on the wrong people too.
I do think anything we do here is inevitably a compromise, I'm hoping we can find one we're all equally, and minimally, unhappy with.
Dave.