My first bet in this post is that XML Schema (XSD) is here to stay for a while. I don’t even think that there is much disagreement about this matter. Regardless of the fact that I am more of a Haskell aficionado than an OO follower, I also bet (in fact, I take for granted) that OO is here to stay as well. (A more risky bet, only mentioned aside, is that perhaps Haskell will become more OOish.) Given the two conservative assumptions about XSD and OO, I wonder: Why isn’t there more research on getting these two to marry?  (Read as in “research on X/O mapping or XML data binding”) May I bet then that establishment “considers XSD as harmful”? If so, perhaps let me continue guessing that the second out of the two choices is predominant: ‘fight the disease or pretend that it does not exist’. Anyway, both options have been neither effective nor constructive in the case of Cobol, and I vehemently bet that they will fail in the case of XSD, too.

 

On a related account: The last year or so I have thought deeply about canonical X-to-O mapping, and I was so fast with getting a handle on the subject because Erik was around. I have a hard time convincing myself that I am as smart as him (still working on that part of my ego); so it is insightful to notice that it took 5 years for Erik to get a full handle on ditto subject. For the readers of this blog, I offer a non-exclusive opportunity: learn most of what I learned on X/O by reading a plain paper; link forthcoming in this post. But please understand – all fun is over once you click that link. I lost a lot of sleep over canonical X-to-O mapping. Now decide for yourself: the blue pill or the red pill. If you opt for the red one, please do send comments which I am happy to take into account for finalizing the draft. Apologies for getting so pathetical: “Please stop XSD bashing”. Research needs to get up to speed and take XML Schema serious. Even if you (not necessarily me) want to transition to a new version of XSD or to a different XML type system, there is no way around it: we need to deeply understand XSD given its level of deployment.

 

Best,

Ralf