I'm quite sympathetic to NVDL; having used it to cleanly integrate XHTML and XForms, it's proven to be a real boon just from that standpoint alone, and I suspect that this integration of disparate (albeit complementary in this case) namespaces will continue to be its bread and butter usage for some time.


However, one thing that I have had painful recent experience with is the degree of investment that the status quo (most notably tool vendors) have in XSDL and the sometimes near panicked response made when Relax NG in particular looks like it may threaten that investment. Relax NG and XSDL overlap over a significant aspect of their domain because they are both grammatical validators, though the fairly stubborn rate at which RNG is becoming the alternative not to XSDL but to DTDs makes me suspect that the language has legs its just superior to XSDL for documents, and despite the rise of SOA, document-centric XML is still the dominant usage of the language.


Validation as a concept is generally poorly understood by the aforementioned mainstream. Schematron in particular illustrates that validation does not have to be complete (i.e., all possible assertions that can be made of a given instance are in fact made), and that there really is no such thing as one unique validation for a given XML instance. XSDL tries to take a monolithic approach and succeeds about 75% of the time, but even the XSDL community is recognizing that there are constraints that simply cannot (and should not) be modeled in the language. (I also keep wondering if Kurt Goedel's theorem isn't just waiting in the wings in all of these discussions of (axiomatic?) assertions).


Kurt Cagle
Managing Editor, xml.com
kurt@o...