[
Lists Home |
Date Index |
Thread Index
]
Bob Foster scripsit:
> True, tho I suspect that the basic algorithm many use is "run it through
> trang." ;-}
Of course. But even trang, concentrated essence of James Clark's brilliance
that it is, can't do more than can be done. If you need to have a WXS
output that's equivalent to your RNG original, then you need to design
that original with an eye to WXS limitations. Otherwise you wind up with
a WXS schema that's looser than the original, which is often sufficient
but other times is not.
The same story applies to DTDs, of course.
> Yes, but looking at the code I noticed MSV creates a list of types collected
> in the course of validating an attribute.
I had forgotten about this.
--
Even a refrigerator can conform to the XML John Cowan
Infoset, as long as it has a door sticker jcowan@reutershealth.com
saying "No information items inside". http://www.reutershealth.com
--Eve Maler http://www.ccil.org/~cowan
|