I think that there needs to be a way to _talk_ about distros consistently;
a distro compatibility taxonomy, if you will.  All kinds of packaging
schemes, subsetting, toolbase (GNU vs traditional Solaris, for example),
minimization, tuning, etc are _possible_, but how much benefit is
there from having grossly incompatible distros targetted at similar
platforms and purposes?

There's a good side to diversity (not the least of which is the ability
to test alternative approaches on a large scale), and also a bad side
(learning curve, compatibility, maintenance, ...).

A consistent language for these matters ought to be the basis for
avoiding those divergences that serve nobody, while encouraging
those that have a demonstrable benefit.

For example: how can one define the different phrases regarding
branding that might be permitted in a widely acceptable manner
unless one can first come up with a widely agreed-upon language in
which those definitions can be expressed?   (i.e. how can one work
towards reaching a point where most of us would be happy with the
process leading to being allowed to call something an OpenSolaris
distro vs a distro based on OpenSolaris technology, etc)

As a hypothetical consumer of a distro, I'd want an environment
that would be conducive to the development of one highly
suitable for my needs.  Nevertheless, I'd probably also want to
be able to leverage as much skill, application portability, and
binary compatibility as possible from there to any future
platform/distro/experience.  Linux has the "pro" side of diversity,
because Linux proper is just the kernel (and realistically the libs
that call directly into the kernel).  But because their distros have
nothing else holding them together, unnecessary differences
proliferate.

It's all about balance: neither tyranny nor anarchy; diversity for
practical purposes, but not as an ideal unto itself.  IMO we can
strike a more pragmatic balance since we have almost an entire
OS, desktop and server environment, and even a number of
applications as the code base that we start from.
--
This message was posted from opensolaris.org

Reply via email to