Re: Draft new DFSG
Joey Hess wrote:
> No it's not. Here are a few reasons we may chose to keep a package out
> of debian even though it meets the dfsg:
> * because it is of very low quality, and there is already a far superior
> alternative that is also free
That's almost a good reason -- but if someone really wants to package
it, do we really want to reject it? Wouldn't it be better just to
assign it a priority of 'extra'?
> * because it implements something inherently unsecure, or is written in
> such an insecure manner that fixing it would require a rewrite
If the packager can't fix the problem, and can't arrange to get a fix
from upstream, then I agree that this would be a good reason.
> * because it is 5 gb in size and will only be used by a few hundred
> people on earth.
Again, if someone wants to package it, do we really want to reject it?
We already need multiple CD support. In the longer term, it might be
worth investigating the idea of creating optional "add-on" CDs with
stuff like this, but until then, I'd be pretty dubious about rejecting
something that someone has taken the time to package.
> All of these seem valid to me. The fact is, no-one in their right mind
> would wnat to maintain such packages for debian, so they never get into
> the distibution in the first place. All free software is *not* in
> debian, and it never will be.
Yes, I think that one of the most important criteria should be whether
someone is willing to package something properly, and is willing to
maintain the package at a minimal level of quality and (especially)
security.
The things that really attracted me to debian in the first place are:
1) it's an extremely high-quality system, 2) it's committed to free
software, 3) it's open to anyone who is willing to devote the time and
effort to get their favorite packages included, and 4) it has a clearly
defined set of guidelines for what can be included and what can't.
If we lose points 3 and 4, I will be very upset. OTOH, I think we
should be prepared for Joey's example of a 5GB useless package. Since
the size of slink already forces us to deal with multiple CDs, I think
its worth discussing the idea of having a core distribution on 1+ CDs,
and at least one CD of optional add-ons (still free, still part of main,
but not part of the core Debian system). For example, stuff that's
currently classified as 'extra' could be declared to be non-core. And
if something comes along that's 5GB and useless to most people, we
classify it as 'extra', and it automatically goes in non-core.
The nice part about this scheme is that it causes minimal disruption to
our existing format. But it's just an idea....
--
Chris Waters xtifr@dsp.net | I have a truly elegant proof of the
or cwaters@systems.DHL.COM | above, but it is too long to fit into
http://www.dsp.net/xtifr | this .signature file.
Reply to: