Cart
[Home] [By Thread] [By Date] [Recent Entries]
From: "Bob Foster" <bob@x...> > Whoops. I got carried away and wrote: > > > ...If you're waiting for a significant > > body of all-Schematron schemas, I would guess it will be a long wait. > > Talk about your flame-bait! If I had a cancel I would use it. My apologies > to Rick and anyone else who got that far into the message. Of course there > is a 'significant body' of all-Schematron schemas. The fact that I prefer to > use Schematron to augment other kinds of schemas is beside the point. Not beside the point at all. We are not starting with a blank slate: Schematron arrives with grammar-based schema languages occupying centre stage, and encounters the widespread, tremendous, brave, pioneering and Pollyannaic hope that we can just extend grammars a little bit more we will reach some satisfactory point. That Schematron will be a layer on top of grammars for the forseeable future is just being realistic, not something to feel apologetic about. Every schema paradigm is can express conveniently a different range of constraints. That should be the starting point of thinking about schema languages, rather than the DBMS vendor's el Sahaf-ism that one schema language is all that anyone needs, as was claimed in the early days of XML Schemas. Just this week I had a support call from a user who said that their schema was OK according to a highly-marketed tool, but failed in ours. I checked and it was a really trivial error (in that tool-- we use Xerces and MSXML which got that particular issue right, which may just be the luck of the draw. The issue was to do with propagating namespace declarations from one schema into an <include>d schema, b.t.w.). I am pretty disturbed that such errors can be still present after all the time that WXS has been out; my advice from last year still seems to hold:-- contracts should specify that a couple of different tools be used to validate data when using WXS. Cheers Rick Jelliffe
|
