RE: [xsl] Features of Saxon vs. Features of the Standard

Subject: RE: [xsl] Features of Saxon vs. Features of the Standard
From: "Michael Kay" <mike@xxxxxxxxxxxx>
Date: Fri, 26 Aug 2005 09:03:39 +0100
> Michael:
> I know Saxon is yours, and also that you have contributed 
> significantly
> to the standards.  Why does Saxon has so many useful features whereas
> the standard has shortcomings in key areas?

Many of the features in the standard are there because one of the products
(often but not always Saxon) put it in experimentally as an extension and it
proved popular. On the other hand, some extensions such as xx:evaluate()
have failed to make it into the specs because there was opposition to it.

It's actually much easier to devise and code up an extension function in
Saxon than to write a spec for it that's sufficiently rigorous to get into
the language specification even, if there is no opposition to the principle
(and there usually are objections to new features, because people want to
keep the language small).

I'd have some reservations about putting things like saxon:transform() in
the spec myself. I think this properly belongs in a pipeline processing
language that's well integrated with XSLT but separate from it. However,
it's easy to provide and convenient for users who want to do everything
within one language, so I put it in - I tend to be pragmatic about such

Michael Kay

Current Thread