Well, my work on implementing an XPath engine within Mandarax is finally looking up. There is only one area that I'm facing big troubles, and I'll start worrying about that later (tomorrow most probably). It's cool what it can do now - you can give a query like "/fact | /rule" to gather all the rules and facts within the knowledge base, or if you want all rules with a certain size, you simply go "/rule[@size='10']". The number of queries is unlimited - I've built a plugin system so people working within Mandarax can easily add further datatypes to the engine, allowing it to then provide results for that data.
Compare this with having to write your own code each time you want to find something out within the knowledge base, and the benefit becomes hugely obvious. And here is a view on the code repository zoomed in on my XPath code (I admit the code isn't nice right now - it's in debug mode):
Other than that, I've just been finalising (and bug fixing) Leopard Tutor, my app for a massey associate professor who wants to use it to teach students Java and object-orientation. They basically build up a UML diagram from a source code listing. Hey Microsoft - wouldn't it be cool to get this switched over to a C# teaching tool? It would take very little work, and it's another way to get students to learn .NET over Java. Ask me for more info 
.
.Cheers,
Jonathan Giles.
Thoughts on “XPath and Leopard”