RE: second implementation of XSLT 2.0?

> Before a W3C Candidate Recommendation advances to Proposed
> Recommendation status, "the Working Group should be able to 
> demonstrate
> two interoperable implementations of each feature."[1] 

I don't know anything that isn't public knowledge, but I think one can make
an educated guess that Oracle and IBM wouldn't have put the amount of energy
into the WG that they have if they hadn't got something up their respective
sleeves.

Note that the "two interoperable implementations" do not have to be released
products.

> I've recently
> learned that XML Spy is advertising XSLT 2.0 support[2]; does anyone
> know if they wrote their own engine and how good it is? 

Yes, they wrote their own engine (or at any rate, if they copied my code,
they introduced some deliberate bugs to hide their traces); no, I don't know
how good it is, other than the anecdotal evidence I get from the occasional
user who tells me when it gives different results from Saxon.

The XSL WG is working on creating a test suite. It's not as far advanced yet
as the published XQuery test suite (and there's no guarantee that it will be
published, since publishing the tests is as much work as developing them).
But it will be used to satisfy the criteria for advancing to a Rec.

Incidentally, the W3C process document cited in the message is an old
version. The current version is

http://www.w3.org/2005/10/Process-20051014/

although the version used to assess these spec's readiness for CR was

http://www.w3.org/2004/02/Process-20040205/

The relevant requirement is now expressed as: [The WG has] Shown that each
feature of the technical report has been implemented. Preferably, the
Working Group SHOULD be able to demonstrate two interoperable
implementations of each feature. More specifically, the criteria for XSLT
2.0 to exit CR are described in the Status section of the document:

<quote>
The agreed exit criteria for this Candidate Recommendation (that is, the
criteria for its becoming a Proposed Recommendation) are as follows:

   1. Sufficient reports of implementation experience have been gathered to
demonstrate that XSLT processors based on the specification are
implementable and have compatible behavior. The WG expects to have at least
two implementations of each required feature.
   2. An implementation report has been produced indicating the results of
testing each distinct feature in the test plan. The WG believes that these
features cover the specification.
</quote>

The phrase "independent implementations" is not used, but I don't think the
Director would be very impressed if the two implemementations were produced
by compiling the same source code with different Java compilers.

Note the insertion of the word "required" in the phrase "two implementations
of each required feature". It would be reasonable to infer that the working
group was not confident that it would be able to demonstrate two
implementations of every optional feature. 

Michael Kay
http://www.saxonica.com/