Re: Q to all candidates: SWOT analysis
On 3/20/19 1:44 PM, Joerg Jaspert wrote:
>> * What are the main 2-4 external things happening in the world outside
>> Debian, and that are "threats" for Debian?
...
> One of the questions in my platform hinted at one point: The "package"
> managers various new languages came up with.
Do you (and other candidates) see this as a threat or as an opportunity?
Instead of packaging libworld+dog (last time I guesstimated it, we had
<10% of Ruby and Perl, <5% of Python and <1% of JS...) do you envision
APT being aware of other package managers (npm, pip, cran, cpan, nix,
snap, something something go, etc.) and being able to say "npm is
outdated, and this isn't the same risk level as having any other
Priority: optional package out-of-date"?
The reason why I think this is a question for everyone and not just APT
maintainers is connected to a question I made in another thread. I asked
whether Debian is more than Debian GNU/Linux the distro you download and
install or that you get from a hoster or that you get as a side effect
of someone choosing FROM debian in a Dockerfile.
I implied that Debian has technical deliverables beyond a distro: APT,
understood broadly as the libraries, frontends, low-level tools, build
and packaging tools, linter and QA, processes, policies and procedures,
the package metadata definitions, etc. I think this has tremendous value
for the industry. There's also of course the packages themselves, all of
which are important but many of which have significantly more logic in
them compared to upstream.
Seen this way, the distro (whether a chroot, a Docker image, a CD) is...
a "reference implementation" of the Debian worldview, and so are the
pure blends. There's a perception that the "distro" is easy to replicate
(very hard to maintain, though...) but the "pieces that make the distro"
are a unique superpower of this project. Furthermore, I think we might
be able to find a lot of energy (researchers, enthusiasts, companies)
that are discouraged from working on "distros" (feel it's not the 90s
anymore, it's passe, etc.) but are really excited about working with all
the underlying technology.
And when coupled with DFSG, I see even the Hacker News crowd looking at
Debian as a bit of an oracle when it comes to this stuff. It might very
well be that we're going in this direction anyway, but I don't think we
are sending that message today (or that we want to)
(Yet another mail that should have gone to -devel...)
Reply to: