On Sat 2021-11-06 11:32:35 +0100, Martin Steigerwald wrote:
Pierre-Elliott Bécue - 06.11.21, 11:06:58 CET:
That being said, the name is indeed outdated, and "Debian Archive
Team" sounds quite nice.
Agreed. I like this name.
Yes, please. "Debian Archive Team" is fine. This is fair amount of
work, but it will help make debian not seem quite as archaic as I'm sure
it seems to new prospective users or developers. Thus it is valuable
work. But a GR does not seem necessary.
The way to do this is to consult with the people already on the team and
the DPL. Select a new name, figure out what work needs to be done
to make the change. Make a plan, have someone to drive it, and
persist. It will probably take at least a full debian release cycle.
Changing names is hard, even if you don't have reactionary pushback.
Many things might be touched by this: e-mail addresses; mailing lists;
text in debian policy or the developer's reference; DNS labels; OpenPGP
certificates; SSH host information; wiki entries; software like dupload;
etc (fortunately, the archaic team name doesn't appear in the
Constitution or the DFSG).
Consider upgrade paths and how to deprecate the old name safely: when
updating e-mail addresses, can you create an alias from the old label to
the new address? How about DNS records? How should we handle mailing
list archives? When/how should you send a deprecation warning when
people use the old label?
Have a timeline that acknowledges the work involved, and plans when to
take each step. For example, changing DNS records, e-mail addresses,
and cryptographic associations will probably be slower/more cumbersome
than changing human-readable labels. Be prepared to revise the workplan
when someone discovers some other place that the old name is embedded.
You need to find someone or someone(s) who have the capacity and the
skills to actually carry out the right work -- or who at least can keep
track of the work and encourage/support the folks who have the
permissions to do it to get it done.
No one should object to this work if it's done with this kind of
thoughtfulness, care, and attention to detail.
Helping the project through this transition would be a great
contribution to Debian, because it fixes a silly stumbling block that
existing developers have already learned how to ignore, but that does
actually hold the project back from welcoming new members who might have
never heard of FTP (or of using the term "master" to mean administrator
for a machine) before.
This work is *not* the kind of contribution that maps cleanly to a
facility at packaging free software for redistribution. This is a great
example of why we need more than just package-maintainers as debian
developers. There are probably many other parts of the project that
need this kind of attention and effort, and we should absolutely *not*
scare people off who want to help fix things.
But let's not make it harder to fix than it already needs to be by
dragging a GR into the mix as well.
--dkg
PS For people who are concerned that a retreat from the term "master" is
somehow the language police run dangerously amok, it's worth asking
why you feel so committed to the term "master" that you would fight
to keep the project we all work on using terminology we all can
acknowledge is confused and outdated. If someone is excited to
improve the project, even if you don't have the capacity to help them
do it, at least *let* them do it!