Mike "Ford" Ditto writes:
> inconvenience.  In fact, I would turn your question around: Why is GLDv3
> being handled *differently* from traditional rules for Solaris?  There

Agreed.  It's a lack of engineering.  If the work were thorough
enough, then there'd be no need to make repeated incompatible changes
in the Nemo framework.

Unfortunately, that's not where we are.  Not only have there been
multiple incompatible changes, but there are more to come, and
seemingly no end in sight.  (We can hope that after Crossbow and
Clearview, we'll be "done," but I guess I'd suggest waiting and
seeing.)

I fully agree with you that this *needs* be a stable interface.
However, I disagree that it's failing to abide by the traditional
rules for Solaris.  In fact, this is *very* traditional.  It's not
stable, so it's not documented and, to the degree that we can manage
it, it's hidden and users are warned away appropriately.  That's
traditional Sun behavior.

Traditional Sun engineering tries to avoid leading users and ISVs and
others off of known cliffs.  That's what you're seeing here.  If you
see the barbed-wire fence and the "turn back now!" signs, but you
drive on anyway, then that's your problem.  We provide GLDv2 and DLPI
as standard and supported interfaces; those are what non-ON drivers
must use until GLDv3 is ready for prime time.

The big problem here is a lack of follow-through, and it's not just
Nemo that's affected.  We have many other famous interfaces (file
systems, anyone?) that are frequently (mis)used by other projects, and
are truly needed by other people outside of ON, but are not actually
stable enough for anyone to use safely.  Thus, they're just land mines
for the unwary.

The continuing crisis is that there's always budget available to add
yet one more feature, but there's almost never time to finish the job
and make it stable.  It takes work to do that, and it's hard work that
doesn't necessarily result in fame and fortune.

> Of course, like many, What I'd most prefer to see is GLDv3 completed and
> made a committed interface.

Yep.

-- 
James Carlson, Solaris Networking              <[EMAIL PROTECTED]>
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
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to