RE: XSLT 2.0 light?

Thanks for the comments. You're not alone in finding XML Schema Part 1
difficult to digest. Clearly in writing specs it's not our job to produce a
tutorial, but we can certainly bear your comments in mind. In particular, I
think we probably need to be clearer about how processing happens in the
absence of a schema: at the moment we rather rely on people knowing that
this is just a special case of the PSVI, and the specs concentrate on
describing the general case which is of course much more complex.

I think some features of the XPath 2.0 type system are things you are just
going to have to live with. The introduction of sequences, I think, brings
many benefits. The expansion from the three scalar types (boolean, double,
and string) of XPath 1.0 to the 19 primitive types of XML Schema is a larger
expansion than many of us would have liked, but many of the primitive types
will hardly ever be used so you will be able to ignore them most of the
time. Complex types arise only when using schemas, and again I hope we can
ensure that when you don't use a schema, you don't really need to be aware
of them.

There will probably be a conformance level in XSLT 2.0 that doesn't require
schema support; defining conformance levels is something we have not yet
tackled.

Michael Kay


> 
> Dear XSLT WG,
> 
> First of all I really appreciate all the hard work put into 
> the XSLT 2.0 
> and related specs. The new grouping facilities look 
> wonderful, and the 
> result tree fragment is a particularly good riddance. I apologize in 
> advance for the whining to come...
> 
> I have one, rather big, problem with the spec in its current form: I 
> don't understand it. Suddenly, to understand XSLT, I have to 
> understand 
> the XML Schema spec (and I'm, regretfully, having a hard time 
> already at 
> http://www.w3.org/TR/xmlschema-1/) and in particular the PSVI, the 
> modified PSVI as defined by the XQuery/XPath data model, the XQuery 
> formal semantics, the multitude of language about types in the XPath 
> specs themselves, and what have you. And I don't, quite frankly.
> 
> That leaves me with little hope of understanding what an 
> arbitrary XSLT 
> 2.0 style sheet actually does - and this bothers me. It may 
> be that I'm 
> unusually slow, or that there will appear some article any 
> day now that 
> will make all the intricacies of the type system clear to slow people 
> like me, but I'm still worried.
> 
> To me, it seems that the step from XSLT 1 to XSLT 2 is way 
> bigger than, 
> say, from K&R C to the latest C++ standard. And that's a 
> rather big step.
> 
> So: Most of all, I would like to see XSLT 2 minus the type system.
> 
> Seeing that this may be unrealistic, I would very much like to see a 
> named conformance level for XSLT/XPath 2 that doesn't involve 
> the type 
> system (in other words: something like XSLT 2 + XPath 1, 
> essentially). I 
> would also like a way to tell the XSLT processor to use ONLY 
> that level, 
> say by means of xsl:version="2.0-light" or somesuch.
> 
> Barring that, an XSLT 1.1 consisting of XSLT 1.0 + grouping + 
> multiple 
> output documents - result tree fragments would be more than 
> great for my 
> concerns.
> 
> Again, thank you for your good work, and for your time,
> 
> /dan
> --
> Dan Holmsand
> dan@eyebee.com
> 
>

Received on Friday, 10 May 2002 09:35:42 UTC