First off, if you ever have to write XSLT, or more importantly teach someone how to write good XSLT, then you should bookmark this article. I especially liked the part where part of his solution to the XSLT problem was to restructure the XML he was transforming. While not always an option it's something I've done many times and it really can help.
You have probably noticed that I've not been posting much of late. Lots going on with the new boss and ongoing recruitment. Over the last few days at work I've had much fun writing some useful code. Of course, as one of my collegues pointed out all of our code is useful, but my point is that over the last three months I've mostly been writing one off Java and XSLT transformation code, probably a couple of thousand lines of XSLT alone.
The last few days I have been re-writing our database creation and default seeding code. Previously we've been using SQLServer specific scripts but it was finally time to bite the bullet and make all of this database independant. Not as big a job as you might think. We have the DataStoreInterface api which should abstracts the SQL into objects. So all I had to do was some up with some XML grammer to represent the table definitions and then to describe the default data to go in them. The only slightly hard part is managing and resolving all the cross-references because of foreign keys. Well it's been a nice diversion from the other work.