Lists Home |
Date Index |
> "support for schema" as per the requirements document should mean
> support for functions over schema datatypes (this is a good thing).
> It shouldn't mean the input is no longer XML but rather the shadowy
> PSVI. Assuming that battle's lost and the input will be a PSVI there
> has to be some way of being sure that when you use document() or the
> implied input document, that you are acting on the document as
> specified. In fact given that the schema processor goes to some lengths
> _not_ to change any propoerties in the infoset of the document, but only
> to augment the infoset with new properties, having XSLT receive the PSVI
> as input would be fine so long as by default XPath operated over the
> original infoset and only used the additional schema properties on
Well, remember, the way it is now, you ***will*** get a PSVI even with
no-schema processing. It's just that all the types will be xs:anyType or
xs:anySimpleType, and all the validation flags will be set to "not
validated". Presumably, that PSVI will closely resemble the infoset you
would get today, if only the XPATH model used the infoset at all.
If the default types are xs:anyType, etc., I wonder how XPATH can be
extended in the future to allow other schema systems, like RELAX NG. Maybe
it won't be so hard to do if the PSVI info includes just the schema
namespace and type names.