Re: X and LSB

[tytso@mit.edu]

   From: Jochem Huhmann <joh@gmx.net>
   Date: 19 Mar 2000 10:02:45 +0100

   But: If the LSB is such vague because it tries to embrace too much, it
   won't be sufficient to test against, so ISVs will take a distribution
   that is "LSB compliant" and use that as test-case. That's in no way
   better than just using Redhat as standard. IMHO there is no other way as
   using existing systems as such a test-case, since ISVs don't produce
   applications for the LSB but for the market. IMHO it is impossible to
   embrace everything ISVs need with such a single specification. It can
   only serve as a common ground to work from and to build other
   specifications on top of that.

This means that the future of what ISV's will code against will be
decided in an open standards forum, instead of dictated by a single
distribution.  I think this is far better letting this be done by a
single company.

It's impossible to embrace the needs of *all* ISV's, but I think we can
address the needs of 80% or more of the ISV's, and for the remaining
20%, their life will be much improved.  Even if we can't embrace
*everything* the ISV's need isn't an excuse not to do anything.

As far as being vague, the specification is in fact quite specific about
what it requires.  As we find other areas where new requirements are
ready for newer classes of applications, then LSB 1.1 or 1.2 or 2.0 will
contain more requirements, and we can nudge LSB-compliant distributions
to move closer together.

This is I think a much more realistic approach than the "big-bang"
approach of trying to define what "Linux" means first.   We don't need
to have a common agreement of what "Linux" means, and whether or not X11
is included, in order to have an LSB-compliant application work on multiple
LSB-compliant distributions.  We just need to agree on what LSB means.
This is what those of us who are working on the specification are
doing.  

						- Ted