Re: Q to all candidates: Universal Operating System
* Lucas Nussbaum <lucas@debian.org> [2019-03-20 10:29]:
> 1) So, if you were asked to write a Social Contract paragraph about our
> universality, defining/outlining both what we aim for, and also maybe
> some limits to that quest for universality, what would it be?
I don't have a paragraph.
But I think it's important to recognize that everything has a cost.
Adding a new architecture requires work from DSA, the release team,
the QA team, etc. Adding a new package makes testing harder, makes
mirroring Debian more expensive, etc.
I believe we've created a culture where people feel they should be
able to do anything in Debian regardless of the cost to others.
I believe we should consider the cost to others more and ask: is it
worth it?
People shouldn't feel "entitled" to add their stuff to Debian; they
should convince people it's worth it. And they have to accept when
others who are going to bear the cost don't agree that it's worth it.
So, yes, we should strive to allow people to do as much as possible,
but not to the detriment of the whole.
> 2) More specifically, if you believe that we should not aim for being
> fully universal, *how* (in terms of decision-making processes) do you
> think that we should draw a line about what's acceptable, for
> example to decide how to cater to the needs of an hypothetical Debian
> GNU/Darwin on m68k port? And what's your own opinion on where that line
> should be (specific examples could rely on debian-ports, release
> architectures, support for non-Linux kernels, init systems, ...)
If something requires changes to every package and they are willing to
do the work and the change is easily maintainable, fine. If they are
not willing to do the work, too bad. If the change increases
complexity, we have to ask: is it worth it?
The release team, DSA and ftp masters already have the authority to
say "no". Maybe we need to support them more when they say "no".
--
Martin Michlmayr
https://www.cyrius.com/
Reply to: