Why does XQuery have to be a "programming language" anyway?
This comment by Mike Champion to Martin Probst “Why do XML APIs Suck” (which, BTW, is a great read in and of its own right) leaves me asking that question.
Don't get me wrong: I agree with both Mike and Martin that something in the way of a full featured programming language for the serious heavy tooling of XML is needed. I just don't see the need to day to try to push the square peg that's XQuery into that round hole. The real value of XQuery today is ability to, in a fairly graspable single step, query and shape XML instances over a set rather than a series. That's hugely valuable to us when compared to the XPath and XSLT approach. Yes, I'd very much like to see XPath with DML too, but not because that completes it as a programming language, rather, because it completes it as a query language.
Its very easy to take what Mike has been saying as of late as painting a gloomy picture of the future of XQuery on the whole. Don't. Understand that what's he's really saying is that he's not sure that XQuery will ever evolve into something more than a Query language. I, for one, don't think it needs to, at least not for a few years.
After all, I don't expect that I'm ever going to write my presentation or business logic code with T-SQL either.