This is interesting, technical work which should be published somewhere. Unfortunately, the link to generative programming or software transformations is a bit too weak. Also, a 2 pages abstract should not be so technical but rather place the work in a context and clarify its originality. Also, the paper should say more clearly why "less recursion" is desirable. Finally, the paper comes a with "universal critique" of XPath papers; that's somewhat harsh and generic for a scientific publication. (There are normally reasonable reasons for abstracting away certain details in papers on XML.)
Well, I agree to it (except the last two sentences). More, with the fresh look I found I mixed at least three things together:
* A cool technical development
* Creating an use case for algebra of programming
* "Universal critique" of XPath papers
Meanwhile, I got a nice question, what is the principal difference of my work and XQuery and XPath semantics. I feel there is difference, but I can't find it. Still thinking.