Re: commitfest.postgresql.org is no longer fit for purpose

On 5/17/24 14:51, Andrey M. Borodin wrote:
>
>
>> On 17 May 2024, at 16:39, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>>
>> I think Andrey Borodin's nearby suggestion of having a separate CfM
>> for each section of the CommitFest does a good job revealing just how
>> bad the current situation is. I agree with him: that would actually
>> work. Asking somebody, for a one-month period, to be responsible for
>> shepherding one-tenth or one-twentieth of the entries in the
>> CommitFest would be a reasonable amount of work for somebody. But we
>> will not find 10 or 20 highly motivated, well-qualified volunteers
>> every other month to do that work;
>
> Why do you think so? Let’s just try to find more CFMs for July.
> When I felt that I’m overwhelmed, I asked for help and Alexander Alekseev promptly agreed. That helped a lot.
> If I was in that position again, I would just ask 10 times on a 1st day :)
>
>> it's a struggle to find one or two
>> highly motivated, well-qualified CommitFest managers, let alone ten or
>> twenty.
>
> Because we are looking for one person to do a job for 10.
>

Yes. It's probably easier to find more CF managers doing much less work.

>> So I think the right interpretation of his comment is that
>> managing the CommitFest has become about an order of magnitude more
>> difficult than what it needs to be for the task to be done well.
>
> Let’s scale the process. Reduce responsibility area of a CFM, define it clearer.
> And maybe even explicitly ask CFM to summarize patch status of each entry at least once a CF.
>

Should it even be up to the CFM to write the summary, or should he/she
be able to request an update from the patch author? Of at least have the
choice to do so.

I think we'll always struggle with the massive threads, because it's
really difficult to find the right balance between brevity and including
all the relevant details. Or rather impossible. I did try writing such
summaries for a couple of my long-running patches, and while it might
have helped, the challenge was to also explain why stuff *not* done in
some alternative way, which is one of the things usually discussed. But
the summary gets very long, because there are many alternatives.

>
> Can I do a small poll among those who is on this thread? Would you
volunteer to summarize a status of 20 patches in July’s CF? 5 each week
or so. One per day.
>

Not sure. For many patches it'll be trivial. And for a bunch it'll be
very very time-consuming.

regards

--
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company