unzoo integration [PSARC/2008/183 FastTrack timeout 03/14/2008]

[James Carlson]

David.Comay at Sun.COM writes:
> I think the interesting architectual question here is the one that
> Darren brought up - is "unzoo" incomplete without the corresponding
> "zoo" component?  Given that the archive format in question is historic
> in nature at this point, I think it is complete since building new
> "zoo" archives seems far less important than being able to extract
> files from old archives.

By ruling out "zoo" but not ruling out "unzoo," it sounds to me like
you're making that argument from the point of view of just one of the
consumers of that package repository.

Isn't it just as plausible that there'd be a "kitchen sink" user
desiring all possible binaries?  Or one who desires symmetry?  Why
can't they get "zoo" if they think they want it?

Darren's question is a good one, I think.  But if the idea is to
create a global repository from which all can choose, it seems to me
to be hard to rule out anything.  On what basis would we do that?

The repository you're talking about doesn't appear to me to be a
system.  The subsets that people choose represent systems, as do the
useful subsets that others might bundle together as a product.  By
selecting things to place into this repository, I don't think we're
necessarily building a system, which means that system architecture is
much harder to see at this level.

It's like looking for "architecture" among the freshmeat.net pile.

-- 
James Carlson, Solaris Networking              <james.d.carlson at sun.com>
Sun Microsystems / 35 Network Drive        71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677