Re: [xml-dev] XPath 2.0 Best Practice Issue: Graceful Degradation

• From: "James Fuller" <james.fuller.2007@...>
• To: "Costello, Roger L." <costello@...>
• Date: Tue, 29 Jan 2008 22:03:22 +0100

On Jan 29, 2008 8:48 PM, Costello, Roger L. <costello@...> wrote:

> 2. Do you agree with all the above?

kind of....

schema validation sometimes gets confused with other things (that we
have present in our coding languages) like;

* try/catch errors
* assertions
* static analysis warnings/errors
* type errors ('surely this is for validation' he says)
* etc

so if one doesn't get confused with the above programmatic structures,
then I do not think there anything  fundamentally wrong with version
1; being able to partially apply schema validation in a procedural
manner must have some uses... though doesn't schema-element() have
something to do with substitution groups ?

All this kind of reminds me of partial xslt processing using that java
juxy thing, which is useful in unit test situations.

I think that there is a problem with version 2 in that it is making a
lot of assumptions with not having any namespace information in the
form of fully qualified elements ... past that version 1 represents a
corner case in xpath2 that could probably be replicated with
combination of xpath1 and other stuff (xsl, xquery) ...  but if its
being argued for schema validation I would just suggest using
schematron, if the itch is so great to use xpath (whatever version) to
constrain/validate one's xml.

cheers, Jim Fuller