2. With all the negative opinions about Linux around here, I'm
surprised
to have to say this, but: Multiple distributions without a
reference for
compatibility is *not* a feature of Linux we want to emulate! I
know, I've
spent the better part of the last 5 years trying to clean up the mess,
with mixed results. It's far easier to create an ecosystem of
compatible
implementations if you *start* with a reference. All attempts at
building a reference after the fact in Linux have been an abject
failure.
Then document some standards, and make that the requirements to carry
some type of identifier indicating compatibility to that standard
(and I do believe there is value in this, and it's probably something
that should be done sooner rather than later, if it is to be done).
But please don't call that standard the "OpenSolaris Standard", and
don't raise one specific distribution to become the "OpenSolaris
Reference Distribution". Either of those implies endorsement by the
community at large, and implies that those that don't comply aren't
really doing it right...
The reason I'm not keen on that is that I'd really like to see
someone build something on top of the OpenSolaris kernel that doesn't
want or need to be like something that's already out there: something
that's not like Solaris, that's not like GNU/Linux, that's not like
BSD,... (Some random examples: a distribution with a non-Unix-y user
land, or one that removes most of the compatibility features -- so
you'd only have SMF to run services (no more rc scripts), that get
their configuration from within SMF (so no more /etc/
banana.conf)...), the barrier for which is raised if it means
fighting against (the perception of) such community endorsement.
Bart
_______________________________________________
opensolaris-discuss mailing list
[email protected]
