Cart
[Home] [By Thread] [By Date] [Recent Entries]
Andrew S. Townley wrote: > I think if the XSD group would recommend what you're proposing, it would > >go a long way to saying that everything doesn't have to fit in the one >box. Separation of concerns is a proven software design principle after >all. > > Oh, we lost the one-size-cannot-fit-all debate before XML Schemas started. Of course, we get the dolorous satisfaction of being proved right afterwards. That issue is dead now. Instead of seeing XSD as a missed opportunity for layering and incrementalism, we need to see XSD like SGML 1986-1996: a large playground being used anarchically by many different individuals, which ended up showing which features could be removed or refactored to other layers to give a more useful playground. And which proved again that underlayered standards are completely implementable: if you don't mind lateness, bugs, partial implementations, self-defeating shortcuts, and user resistance. On Schematorn, the XSD WG are mostly the right age to recall Monty Python's Mr Creosote sketch (the enormous fat man who explodes with "just a little wafer" more), as well as the knights who say NIH. Cheers Rick Jelliffe
|
