Re: [xml-dev] Adam Bosworth on XML and W3C

[Jonathan Robie <jonathan.robie@softwareag.com>]

At 04:07 PM 10/9/2001 -0700, Ronald Bourret wrote:

>This is interesting. Has anybody else asked for factoring in XQuery? Is
>there a reasonable factoring? (One obvious factoring is to move
>user-defined functions to a separate module, but I don't know the spec
>well enough to suggest others.)

There is, of course, the obvious factoring due to the fact that XQuery 
contains XPath 2.0. There is one class of applications that gets by very 
nicely with just XPath 2.0, and another that requires all of XQuery.

By the way, If I follow the argument Adam is making, the following 
paragraph is central:

 > But the main uses of XML have turned out to be integrating
 > information through coarse-grained messages often stored in XML
 > stores, and for describing the persistent state of
 > objects. Furthermore, it is largely used to access information
 > functionally, not through a data model. SOAP, WSDL, and UDDI are all
 > part of a functional model for retrieving information much like stored
 > procedures in a database rather than a data model such as SQL in a
 > database.

Adam says quite clearly that the main thing he is looking for is an 
efficient way to filter data in messages. That's not the main problem that 
XML Query was designed to address. If I had to guess, I would not be 
surprised if Adam is involved in writing or promoting a language designed 
for just this purpose. But all I have to go on is reading between the lines 
of his article.

 > The original ambitious ideas that XML could be used as a view
 > architecture allowing intelligent data searching across seas of
 > applications has yet to bear fruit.

That's because XQuery implementations and data integration architectures 
designed for XQuery are not yet part of the commercial landscape. I don't 
think that this vision can succeed on the basis of the DOM, XSL, and XPath, 
and I think that XQuery is one of the first standards that is really 
positioned to make this possible.

The conclusion Adam seems to draw is that we should abandon the attempt to 
seriously query XML, and write a simple language for filtering XML 
messages. Maybe a language along those lines would be useful, but XQuery 
was designed to address a much broader set of use cases.

Jonathan