[
Lists Home |
Date Index |
Thread Index
]
At 07:52 AM 1/7/2002 -0500, Jonathan Borden wrote:
>Perhaps the argument can be made not for unifying the 'surface syntax' of
>XSLT and XQuery, but rather the underlying model. To the extend that the
>XQuery formalism is a good model of XML, and incorporates, or references
>good models of XML in general, one would wonder why XSLT 2.0 would need a
>different formalism.
>
>To use your example, this would be analogous to having C++ compile to the
>Java VM, or essentially what MS has done with C#. Hence I say that the
>proper comparison would be between C# and Java -- well assuming a whole lot
>of extraneous issues related to MS and Sun didn't exist :-))
I agree that XPath and XQuery should use the same underlying model, and I
think they already do. They use the same data model, the same type system,
and are even generated from the same grammar. They are formalized using the
same formal semantics. The looser typing of XPath is expressed by fallback
conversions that are not supported in XQuery.
Are there further steps that we should take in this direction?
>That said, if there are some simple changes which would unify the (non-XML)
>attribute constructor _syntax_ that would be a Good Thing.
Agreed.
Jonathan
|
