RE: TR/xquery-operators/#func-doc

> Thanks for the clarification.
> Is this one more case where the needs of xquery and 
> xslt+xpath are 'slightly different'?

No, I don't think that's true in this case. I think fn:doc() is the function
that we would offer XSLT users if there were no need for backwards
compatibility. The only difference between XSLT 2.0 and XQuery requirements
here is that XSLT 2.0 has a requirement to be backwards compatible with XSLT
1.0, and XQuery doesn't.

> 
> As the detail comes out, the number of places where this is 
> occurring appears to grow.

My own perception is that the languages are tending to converge rather than
diverge. For example, the introduction of sequence construction semantics in
the XSLT draft of 2nd May brings the semantics of tree construction in the
two languages much closer together. And of course the "promotion" of XSLT
serialization as a shared specification is another big step towards
convergence. XQuery in its latest draft has also adopted a number of other
established XSLT ideas, for example global and external variables
(equivalent to XSLT stylesheet parameters). We are also talking about moving
XSLT's format-number() into the shared space, though there is still design
work needed to achieve this.

> 
> Does anyone on the WG have a count of these differences,
> where the commonality of requirements do not fit?
>   I would like a response to this question please.
>
The answer to the question is no, no-one has attempted to quantify this. I
don't know how you would measure it. 
> 
> 
> I'll repeat my request.
>   I would like to see the common items common, not nearly 
> common. Then xquery could go do its typed thing to its hearts content.
> 
The two working groups strongly share this goal of making as many things as
possible common.

I'm confused by your final sentence, though. This seems to suggest that you
would like aspects of the type system to differ between the two languages.
We want the type systems to be identical: this is a prerequisite for sharing
other parts of the functionality, and also promises significant user
benefits in allowing XQuery and XSLT to interoperate in applications.

Michael Kay

Received on Tuesday, 13 May 2003 20:30:39 UTC