[sc34wg3] Re: Backwards Compatability WAS: Public Interest and ISO WAS: [topicmapmail] <mergeMap> questions

Murray Altheim sc34wg3@isotopicmaps.org
Fri, 19 Oct 2001 10:25:53 -0700

Kal Ahmed wrote:
> At 13:02 18/10/2001 -0700, Sam Hunting wrote:
> >[kal ahmed]
> > > Yes I would support it. However, your definition of what constitutes
> > > broken and mine are possibly different.
> >
> >Probably, but at least we have agreement in principle.
> Its just a shame that the basis of that principle is not agreed on ;-)
> [Lots of discussion about W3C specs between two people who don't know the
> history of W3C specs deleted...Murray help us please!]

I think leaning on the RDF specification as an example of W3C specs
changing is really stretching it. It's so broken and unloved that its
syntax must change for the base technology (graphs-in-XML) to survive.
The only analogue to quickly-changing specs is HTML, where we saw the
IETF's HTML 2.0, W3C's HTML 3.2 and 4.0, and now XHTML, but there were 
quite a number of years between them and I hope I don't need to repeat 
the history of why these changes occurred. I'd not blame the W3C on that

As for XML specs, I know of *no* W3C Recommendation that's been changed,
other than XML 1.0, which went into 2nd edition (not 1.1 or 2.0) by 
incorporating the known errata and clearing up some parts of the prose
that were unclear (such as the common misinterpretation of xml:space
declarations because of an example). There were no changes to XML due
to this edition and XML itself remains stable (thank god).

This is not to say that the world of XML has been stable. The W3C
Recommendations tend to bounce off of each other, so while XML 1.0
hasn't changed, the Namespaces in XML and other Recommendations have
added layers of complexity and non-interoperability. That and with the
continuing discussions on namespaces, URIs, and interpretation, one
can only be guaranteed interoperability with another vendor's products
who have implemented the same *set* of XML specs, and in the same way.
It's a real mess. I don't think we want to emulate that. 

Jon Bosak, Eve Maler, and others here at Sun have requested a 
moratorium on changes in the XML world to allow the industry to
stabilize around it. While the W3C hasn't responded to that plea,
a similar moratorium might be an idea to emulate. 

Sam, anytime one asks for stability of the XTM spec you respond by
claiming we're thinking of the spec as sacrosanct. Damn right! As
Kal has said, this community has built a small number of vendors
who've gone out and done some business, all around XTM. Even if 
XTM has some flaws, we had better be really sure any changes to
XTM at this point are *reaalllllly* warranted (and not simply a
philosophical difference) because screwing with the vendor and
customer base will certainly kill this nascent industry. Even if
it's flawed (which all specs are) so long as there is functionality
and interoperability, then confidence in the technology will grow.
Making some fundamental changes at this point should be considered
a very dangerous activity, the kind of thing that ends the rest of
these discussions since nobody cares anymore and has moved on to
other things, hopefully more stable things. 

And I hope we can refrain from sky-is-falling rhetoric in discussing
the necessity for these changes. As I *think* is clear, nobody buys
the rhetoric anyway. Technical discussions don't move forward under
such talk.

The best thing that can happen right now is for XTM 1.0 to remain a
stable specification that can allow confidence to be gained in it,
so that the community of vendors, developers, customers and users 
continues to grow. 


Murray Altheim                         <mailto:murray.altheim&#x40;sun.com>
XML Technology Center, Java and XML Software
Sun Microsystems, Inc., MS MPK17-102, 1601 Willow Rd., Menlo Park, CA 94025

               Rally against the evils of iceburg lettuce! 
            Grab a kitchen knife and join the Balsamic Jihad!