RE: [xsl] XSLT 2.0 Vs XSLT 1.0

Subject: RE: [xsl] XSLT 2.0 Vs XSLT 1.0
From: "Michael Kay" <mike@xxxxxxxxxxxx>
Date: Thu, 11 May 2006 18:01:23 +0100
> Thank you for your response. I was reading that XSLT2.0 is 
> still not stable. What are your comments on this? 

XSLT 2.0 is now a candidate recommendation (and has been for six months).
You can form your own judgement of its stability by looking in the public
bugzilla database which records all comments and changes made since the CR
release: go to

and searching for product=XQuery/XPath/XSLT, component=XSLT 2.0,
version=Candidate Recommendation, status=all

or for a canned search (this might not survive the mailer),

This shows a list of 57 issues raised - which sounds significant, until you
look at the detail. They're almost all either editorial points asking for
greater clarity in the exposition, or very obscure corner cases in the
specification, for example removing overlaps between error codes. There have
been no changes in the syntax of the language, and I think it's unlikely
that there will be.

The majority of issues are being raised by members of the working group plus
a few outsiders who (one suspects) are implementing the specification and
finding issues as they develop and test their implementations. There's a
test suite being developed within the WG and this is finding gaps both in
the specification and in products, but it's all at the level of minor

Of course, different user organizations have different attitudes to risk.
FWIW, I know of at least two investment banks that are putting millions of
dollars of transactions through Saxon-SA on a daily basis. But I don't blame
anyone who wants to be cautious, given that there are no products yet from
big players like IBM, Microsoft, or Oracle.

Michael Kay

Current Thread