Kevin J. Jones

Just before the christmas break we released a beta of the new Intel XSLT 2.0 processor. This has been one of the major developments running in my group last year. You can find it over here.

The processor is really just a part of the SOA Expressway XML software stack that we have unbundled to allow people to play/test with. There is no intention at the moment to turn this into a publically available standalone product, the market for such is simply not viable for large commercial vendors. Michael Kay talks about how XSLT 2.0 was just too big for the “weekend coders” to see many open source implementations of XSLT 2.0 emerge. I don’t think you can argue with that one although I might have liked to have tried for the hell of it if I wasn’t busy with other things. That’s how I got started in XSLT/XML development some nine years ago now, by just getting on with it.

We have plenty of work left to do on XSLT 2.0 to create a really great feature for SOA Expressway. Conformance today is very good but the performance needs some extra work. This processor has to hit some pretty high targets to match our tuned XSLT 1.0 performance levels. We have talked about if it makes sense to add XQuery, that’s not such a big development but we have not concluded it’s needed yet. The processor is not yet schema aware although most of the plumbing is in place to make that happen so I would hope we can complete that soon.

It’s hard in software to ever say you are done, but with our XML Software we are getting closer and closer. The benchmarks measured in cycles, memory & scalability have shown improvements year over year to the point of having little serious competion on technical merit and the feature gaps have shrunk very small. My one last goal is to see more profitable use being made of it, I think we are well positioned to get that this year as well.