Hi Francisco, Erik, On Sat, 2026-11-14 at 00:00 +0000, Francisco M Neto wrote: > Em sex, 2026-11-13 às 17:20 +0100, Erik Huelsmann escreveu: > > On Fri, Nov 13, 2026 at 8:35 AM Maytham Alsudany <[email protected]> > > wrote: > > > The problem is that many people treat Debian like a PPA, where > > > they either attempt to upload some low-quality program, or will do > > > a great work packaging but then disappear off the radar, leaving > > > teams like MIA to clean up after them after a long, long time. > > > There is no way to tell this just by looking at an RFS, other than > > > the contributor's previous history. > > I could not disagree more with that take. People are not told > what is expected of them in a clear way; also, there is no distinction > between those who just want their software to be available on Debian, > and those who *actually* want to become a DM/DD. There should be.
Which is why I'm saying it's difficult to decide whether to sponsor an
RFS or not. I'm not saying that "every sponsee needs contribution
history", which is impossible, but I mean that we can't predict the
future, and no-one wants to make the wrong decision and waste time by
sponsoring someone who will end up abandoning the package later.
This is important considering that I as a sponsor have to spend a
non-trivial amount of time reviewing every part of the package for
problems, *especially* when it's NEW and the d/copyright will be
scrutinized, as any problems that get through will end up reflecting on
my ability to properly review as well. When I do an upload, I am signing
off on the work with my own GPG key and saying "yes, this should go in
the archive and be installed on people's computers." If the package is
abandoned, there will also be time and effort spent either removing the
package or finding a new maintainer.
> As for packages being left to the MIA team: I don't think they
> should be picking up every package that is left behind. If taking a
They are not, the MIA team responds to concerns raised by others to
them. Many of these are about sponsored contributors though, which is
why a shortened process for dealing with it was implemented, and MIA is
usually the only ones dealing with this kind of issue.
> certain package out would break things, that's fine. If not, there is
> no point; "having the most packages" means nothing as it regards
> quality. If dropping an orphaned package would have little to no
> consequences, then it should be dropped; instead what I see are
> packages that haven't been updated in years still being propagated to
> new releases, (hyperbole) even if they have 3 installs (/hyperbole).
This is a problem with Debian in general, where people are afraid to RM
a package so they orphan it instead, and then no-one touches it for
years. People working on package QA RM these when they encounter them,
but any mass removal of ancient orphaned packages is likely to start
some toxic debate in debian-devel about "keeping historical packages"
and "it's still useful" and what-not.
> > How is a prospect supposed to get that previous history? I mean: a
> > first contribution will *always* be without previous history, after
> > all...
I should also mention that, for me personally when I am looking through
RFS bugs, there are a handful of indicators that signal whether someone
is willing to actually maintain the package long-term. I don't know how
accurate they are, but it's the only heuristic I can use. It includes
stuff like:
- Properly formatted ITP and RFS bugs, with correct name
- Packaged in a Git repo
- Packaged in a team (if applicable) - much lower chance of abandonment
- Upstream != downstream
- Notable userbase (usually based off GitHub stars if available)
- Genuinely useful
- Updates to existing packages & ITAs have greater value than ITPs
- If outside a team, am I interested enough in the package to keep tabs
on it or add myself to Uploaders?
The Developer's Reference puts all of this into much better words than I
can, quoting below from
https://www.debian.org/doc/manuals/developers-reference/beyond-pkging.en.html#sponsoring-packages
> There's no simple rule to answer this question; it can depend on many
> factors: is the upstream codebase mature and not full of security
> holes? Are there pre-existing packages that can do the same task and
> how do they compare to this new package? Has the new package been
> requested by users and how large is the user base? How active are the
> upstream developers?
>
> You should also ensure that the prospective maintainer is going to be
> a good maintainer. Do they already have some experience with other
> packages? If yes, are they doing a good job with them (check out some
> bugs)? Are they familiar with the package and its programming
> language? Do they have the skills needed for this package? If not, are
> they able to learn them?
>
> It's also a good idea to know where they stand with respect to Debian:
> do they agree with Debian's philosophy and do they intend to join
> Debian? Given how easy it is to become a Debian Member, you might want
> to only sponsor people who plan to join. That way you know from the
> start that you won't have to act as a sponsor indefinitely.
--
Maytham
signature.asc
Description: This is a digitally signed message part

