[
Lists Home |
Date Index |
Thread Index
]
On Tue, 06 May 2003 18:26:56 +0100, Dave Pawson <dpawson@nildram.co.uk>
wrote:
>> If there were one spec defined solely in terms of
>> elements, attributes and text, and another built on
>> top of that one that added typed data, then sure,
>> I'd use the former and simply not bother with the latter.
That's the point of the "conformance levels", so people who just need
elements and attributes and text could (in principle) just use tools that
don't bother with the latter. But as I understand the specs and the
explanation by Michael Kay, the "typed data" is baked into the lowest
conformance level.
"a Basic XSLT Processor must be able to manipulate atomic values conforming
to any of the XML Schema built-in types, for example strings, integers,
decimals, doubles, dates, times, QNames. But a Basic XSLT processor does
not support type annotations on nodes in the data model: all nodes are
untyped. And it does not support user-defined types."
This stuff is in Last Call, so people who believe that this is
inappropriate have until June 30 to make your opinions known. I
respectfully disagree with Joe English that this will just make work for
the WG and they will do what they are going to do irrespective of what Last
Call reviews say; this is HARD TO JUSTIFY in the W3C as it works TODAY.
Unless, of course, people don't submit formal comments, in which case it is
very easy to say "it's just an insignificant minority who want that, we
will ignore them." I was just on a W3C Chairs telcon today, and while
there is certainly a point of view that informed users should have read and
commmented on drafts long before Last Call, there is an equally strong body
of opinion that says Last Call is telling the world, "we're done
experimenting and arguing, now it's safe to take a close look because we
REALLY want your opinion on this."
Sorry to sound like a broken record on this point; I understand why people
are cynical about the Last Call review process, but I think the W3C has
learned from bitter experience that it's important to not ride roughshod
over real, reasoned dissent even if members of a WG want to "just ship it."
Witness the addition of the "web method" feature to SOAP 1.2 after strong
dissent (mostly outside the XMLP WG) by the RESTifarians. Of course, the
result may not make opponents of typed XML any happier than SOAP 1.2 made
the RESTifarians, but that's life in the land of consensus politics.
|