[
Lists Home |
Date Index |
Thread Index
]
Bob Foster wrote:
> Oh, is that the point the REC is very clear on? ;-}
Perhaps Henry is using "clear" in the sense that a needle in a haystack
is clear
when you sit on it :-)
But the root problem isn't the needle nor the organization of the haystack,
it is the size of the haystack (i.e., the technology's size hinders
efforts to fully
implement it, defeating the goal of certainty which should be the
primary goal
of a schema language; especially because it is under-layered.)
I don't want to be a broken record, but it takes six years for software
to become
mature; so, for libraries and applications, XML is mature, XML
Schemas still
has 2 years to go, and XQuery has maybe 7 years to go. Caveat emptor:
when you use semi-mature technologies you will have problems that you
wouldn't (or shouldn't) have with mature technologies. Even if XML Schemas
were layered properly and more explicable, we still would be having a
certain
amount of interoperability issues just attributable to immaturity. The
same issue
affects RELAX NG, Schematron, ATMs, and any software: any "I told you
so" feeling from those of us who complained about the size of XML Schemas
(or its goals) early and often should be tempered by that. The real
test will
come in a couple of years time: if there are still interop problems by
the time
XML Schemas software should be mature, I think it would clearly show
that the technology is intrinsically flawed, at least as far as layering
goes,
and in need of a thorough revision or refactoring. Until then, trying to
get
a Databinder's Profile (Common XML Schemas?) is probably more rational.
Cheers
Rick Jelliffe
|
