Jiri Sasek writes:
> Exactly... The problem is the gnutls is the part of the JDS even if it 
> delivers more core functionality than the "window painting". What more.... it 
> is "internal closed" interface even the "external" is more probably better 
> expressing the position of this library in "opensource world". This is 
> ambiguous behavior of the ARCs when they allow such "closing" of the 
> external-public interface but they does not pushing such "building stones" to 
> provide better stability to the rest of the ?system?

It's a difficult balance to strike.

On the one hand, yes, we in the ARC would very much like to see more
of the useful interfaces being made Committed so that they can be used
by other projects without fear that the sand will shift under them.

At the same time, we cannot in all cases _force_ a project team to
commit resources it does not have.  Making a commitment to a higher
stability level means doing more work: making sure that future changes
are understood and made in a way that preserves compatibility.

We do encourage project teams to step up to what's needed.  We do
provide feedback to project teams and management when private
interfaces are causing problems.  And we even occasionally hold a
project hostage to a stability commitment where necessary.

Part of the problem here is that we're working with a different set of
rules than much of the open source world uses.  For most
incompatibility problems, the open source answer is "download the
newest stuff from our CVS server, recompile, and it should work."
That's great for uncomplicated projects and ones where latest-and-
greatest is exactly what's wanted.  It's not so good an answer if
you're trying to engineer a system where old software still works
tomorrow.

> Being sarcastic I can say ...doing the system means harmonizing the 
> components to work together  ...not make the "consolidation closed" 
> interfaces.  :->

Agreed.

The common alternative to those private interfaces, though, is
documenting things that may be changed radically in the future, and
thus cause applications to fail when the changes do happen.  That's
not good.

-- 
James Carlson, Solaris Networking              <[EMAIL PROTECTED]>
Sun Microsystems / 1 Network Drive         71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to