Re: [xml-dev] XSLT 2.0 / XPath 2.0 - Assumptions

[AndrewWatt2000@aol.com]

Steve,

Thanks for the prompt reply.

In a message dated 12/05/02 15:34:48 GMT Daylight Time, Steve.Muench@oracle.com writes:


> | My guess is that many in the XPath / XSLT community are, at best, only 
> | vaguely aware of the additional complexity that is potentially being imposed 
> | on XSLT 2.0 / XPath 2.0 in the interests of "XSLT as SQL".
>
> Oracle is not one of the companies that is trying to
> push XSLT to be like SQL. 


Thanks for the clarification.

So the "push" for that is coming from elsewhere.

Jonathan Robie wrote:
"The main reason that has been given for including all these features in 
XPath is the claim that XSL users really want them."

To the best of my knowledge "XSL users" have never been asked what they want from XSLT 2.0. I certainly don't recall seeing a request for comments on the XSL list.

I will ask Jonathan again, who made the claim and on what basis?

> I think the features that XSLT 1.0 users want in XSLT 2.0
> are those features that make it an even better "report writer"
> kind of tool. 


Yes, I think that is a reasonable term.

>One of the features I was happy to have 
>worked on was the new XSLT 2.0 grouping facilities. To me,
>while working on designing that feature, I had the strong
>feeling that "users were going to love this!" :-)

That I can see too. :)

>I think the spirit of the W3C to combine the XPath 2.0 and
>XQuery 1.0 efforts was based on the fact that if a user 
>was going to end up using XQuery and XSLT together in this
>way, that it would be quite odd if they used to different
>syntaxes to accomplish the same thing. Prior to unifying
>the core of XQuery with XPath, there was really a lot of
>syntactic overlap. 

Yes, I can see that argument.

What the WG appears to have overlooked is the needs of those who currently use XSLT and XPath for fairly simple "report writing" or formatting.

A significant number of users of XSLT 1.0 and XPath 1.0 don't have a formal CS background. I see these technologies (XSLT and XPath) as an important bridge in encouraging "Web developers" into the XML fold. If XSLT 2.0 and XPath 2.0 hit these guys and guyesses with all the extraneous baggage then the W3C will, in my view, have shot itself in the foot again. A lot of "Web developers" view XHTML 1.0 with derision. It seems to me to be far better to keep these guys onside than push them away from XML.

I found it amusing today to receive a reply on the public-qt-comments list, "... in my view a normative spec should be written to be readable on its own by anyone with a computer science education". Perhaps, of very few readers of that list, I recalled that the same person had written, only a few months ago about XSLT 1.0, ".... the official W3C specification itself ... is about as readable for most programmers as a piece of tax legislation". If the problems in communication arising from XSLT 1.0 were recognised it is very surprising that the corresponding problems with XQuery seem not to be.

Anyway, I am glad to hear that Oracle isn't supporting the bloated XSLT 2.0/XPath 2.0 scenario. Maybe Jonathan or Michael will yet shed some light on who is.

Andrew Watt