Re: About the recent DD retirements
Gunnar Wolf <gwolf@gwolf.org> writes:
> And... Sure, Debian's attractiveness has also morphed. Those of us who
> joined a long time ago (I'm "younger" than you on the project, only
> since 2003, but it's still a very long time) have changed our life
> circumstances, possibly our interests, maybe even our ideological
> viewpoints. And yes, maybe (but that'd fuel a different discussion)
> Debian is less attractive in general to the young developer population
> to what it was in the past — I don't remember where I read that the
> median birth year of DDs has remained almost constant, which means that
> (yes) we might be attracting more senior developers (after all, Linux is
> no longer just a toy), but also... That we are failing to attract young
> talent.
Partly this is because we've been so successful.
Back when I started working on this stuff in the mid 1990s, packaging and
distributing software was a hard problem. The existing solutions were
rather dubious, portability to the various UNIXes was quite challenging,
and we all didn't really know what we were doing.
This situation has improved *radically*, largely (although not
exclusively) due to Debian. The world is now full of high-quality
packaged software, and even the relatively crappy packaged software or
half-assed packages (fpm with no attempt at metadata, for instance) are
better than the state of the art was in the 1990s.
One side effect of this is that packaging feels like a solved problem.
That doesn't mean people aren't willing to work on it, but it isn't the
cutting edge of what we used to call sysadmining and what's now called
SRE. In talking to professional colleagues about this, most of them are
not particularly interested in packaging, not because they think it's
unimportant, but for the same reason why they aren't interested in
coreutils. cp and ls just work, which is great, but it also means they've
never felt any particular desire to work on them.
Addresssing this problem is tricky, since it's a *good* thing that we've
solved a problem reasonably well and packaging is no longer what's
actively painful for people (and thus motivating them to do something
about it). But there are still a lot of hard and interesting problems,
which are worth working on. So, one, we need to figure out how to make
those problems apparent and provide people leverage to work on them. But
we should also consider that the project has a lot of appeal for people
who *aren't* interested in being on the cutting edge of something, and who
find maintaining their corner of a well-established infrastructure rather
attractive. Which requires a different kind of recruiting.
But some of it is also inevitable. A lot of smart, talented people want
to work on really hard problems that are currently major pain points,
since that's where they can have the most leverage. Packaging just
*isn't* that any more, so it's to be expected that a lot of those people
are off working on something else that is: crypto, or social networks that
aren't driven by advertising and surveillance, or container deployment, or
trying to build a programming language that can finally replace C as a
safer language for writing our core infrastructure, to list a few obvious
examples.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: