Re: [xsl] Is there a reason for not using XSLT 2.0 as a default

Subject: Re: [xsl] Is there a reason for not using XSLT 2.0 as a default
From: Wendell Piez <wapiez@xxxxxxxxxxxxxxxx>
Date: Wed, 09 Mar 2005 12:30:36 -0500
Hi Mike,

At 11:02 AM 3/9/2005, you wrote:
I can only speak to the *current* perception in the WebData XML team
at MS about the lessons we as a company and an industry learned from
this experience. The sense I get from my colleagues who were around is
that it *was* a good faith effort to implement what they understood to
be the draft spec, along with various improvements to make it suitable
to known customer needs.

Arguing over whether the effort was in good faith isn't and wasn't very germane. Like you, I assumed the effort was good-faith -- which only made me more disappointed when I discovered what the implementation actually did. In particular, the "various improvements to make it suitable to known customer needs" were problematic, as they were impossible to disentangle from the parts that had once been a WD. (And this was after engineers at MS had helped to define the namespace mechanism.) The upshot was that from the point of view of developers to whom standards conformance mattered -- and not even for the portability of the code (at that early point), but only for the applicability of the knowledge we were acquiring -- the implementation was sadly crippled. To rely on it at all seemed like going to France to try and learn Italian. Not, by definition, impossible -- it's true French is somewhat like Italian, as languages go: but one could also stay in New York, or even download a copy of XT and go to Italy.

  Furthermore, it had the result of offering considerable
credibility to XSLT and creating a demand for XSLT tools and
experience..  I can very easily imagine a world in which XSLT shared
the fate of XLink, if  MS had waited for the final spec and for
customer demand to emerge before supporting it in its core products.

It's generous of you to think so, but I think this is overstating the case. XLink is an entirely different kettle of fish (it wasn't and isn't very clear where it belongs: it perhaps should have been part of a modular HTML+++, or of XSL-FO) and didn't have an XT -- or an (albeit small) community of DSSSL users eager to put it through its paces. The usefulness of XSLT was dazzlingly clear, and web sites with XT behind them appeared almost immediately.

But why is this relevant? The motives of MS back in 1999, or the effects of the decisions they made then -- all we can do is learn from experience. Experience teaches us "caveat emptor" -- even when we have little reason to believe the vendor is not acting in good faith.

The MS position going forward is, as I understand it from my rather
brief experience, NEVER AGAIN -- we will not ship support of a draft
Recommendation in actual products.

Probably as wise now as it would have been then.

  That is why we removed the preview
implementation of XQuery from the .NET 2.0 framework, that is why we
are waiting until XSLT 2.0 is actually a Recommendation before
announcing any implementation plans or schedule. (XQuery  in SQL
Server is a bit of a special case ... in any event we're not claiming
to ship a conformant implementation, just something that leverages the
years of experience that have gone into XQuery and meets pressing
customer needs).

I hope it's not being labeled "XQuery" then (if it's not conformant, it's not XQuery, right?) ... after having resolved not to make the same mistake twice. :->


====================================================================== Wendell Piez mailto:wapiez@xxxxxxxxxxxxxxxx Mulberry Technologies, Inc. 17 West Jefferson Street Direct Phone: 301/315-9635 Suite 207 Phone: 301/315-9631 Rockville, MD 20850 Fax: 301/315-8285 ---------------------------------------------------------------------- Mulberry Technologies: A Consultancy Specializing in SGML and XML ======================================================================

Current Thread