| |||||||||||||
|
| |||||||||||||
Distros, not ISVs, more deserve help -- and can't standardize w/o itDistros, not ISVs, more deserve help -- and can't standardize w/o itPosted Aug 1, 2008 16:41 UTC (Fri) by krishna (guest, #24080)In reply to: Distros, not ISVs, more deserve some help by drag Parent article: Will LSB 4 Standardize Linux? (InternetNews)
<quote>
The only reason Ubuntu works as well as it does is because of the massive
amount of personal effort Debian developers put into making Debian
packages work, and even then because Ubuntu doesn't care so much about
package compatibility they screw things up so that their 'stable' system
is consistently buggier then Debian's 'testing'.
</quote>
Which is one reason Debian's Filesystem Hierarchy Standard and packaging
requirements are IMHO the gold [meta-]standard for standardization. I
think they got it right when they produced lintian, and flag installation
and interoperability problems as show-stoppers for any given package -- in
a way, it's like the US Armed Forces' boot camp -- you come in a tar.gz,
they strip you down, issue you a uniform (the FHS), teach you the rules
(the control file), and run you through a gauntlet until you either drop
out and are not included in the release, or conform and can be guaranteed
to work with others.
As a result, you do have packages that are almost totally guaranteed to
work together, with 'almost' being well-defined by a rigorously specified
set of requirements. They take interoperability and standardization as
seriously as everything else should be taken in software, and it shows in
the conspicuous lack of installation problems in Debian's (albeit delayed)
releases -- when a package is told to 'install', it knows what to pull in,
can count on everything it pulls in following the same rules, and the
installation succeeds.
Assembling a standard for distros that don't believe making
standardization compatibility an absolute requirement of the distribution
qa process seems to provide a tool while sidestepping the attitude towards
standardization.
(Log in to post comments)
Distros, not ISVs, more deserve help -- and can't standardize w/o it Posted Aug 1, 2008 19:51 UTC (Fri) by drag (subscriber, #31333) [Link] If it was up to me and I was the king of the universe I'd issue a decree that everybody should simply drop what they are doing and re-base off of Debian Testing. _Right_Now_. Debian has the most complete, most consistent, and most bug-free packages anywhere I've seen in any Linux distribution, bar none. And this is due entirely to the sort of thing you've described AND the tremendous amount of work and organizational effort that they put into it. Ubuntu _almost_ gets it right. Their mistake was not working towards backward compatibility. If everything in Ubuntu was installable in Debian and visa versa that would save everybody involved a lot of effort and a lot of heartache. IMO, Ubuntu should be not a entirely new distribution, but a installer and a repository based off of Debian. A carefully tested and configured default configuration that differs from what you'd get from Debian's "tasksel desktop". That is you would end up with something like: deb http://example.com/debian/ Hardy main contrib non-free ubuntu One distribution that seems to get it right is 64Studio. They do a very good job of making their packages work with Debian transparently, while still making significant changes to the kernel and other important system files. ------- Of course this isn't going to happen and I am sure it would be a unpopular concept among non-Debian developers.
Distros, not ISVs, more deserve help -- and can't standardize w/o it Posted Aug 1, 2008 21:05 UTC (Fri) by michaeljt (subscriber, #39183) [Link] Ubuntu's marketing is based around being something different and special. I also often feel that they stick too many of their limited resources into reinventing wheels their own way, rather than reusing other people's work well.
|
|
||||||||||||