- From: <noah_mendelsohn@us.ibm.com>
- Date: Tue, 12 Feb 2002 10:29:09 -0500
- To: kellyly@hotmail.com
- Cc: xmlschema-dev@w3.org
Kelly Lynch writes: >> From my reading of the XML schema datatypes, I would expect that >> the datatype value is supposed to be formatted in the "canonical >> lexical representation". However, I've found XML validation tools >> (namely MSXMLDOM 4.0) that enforce the less restrictive "lexical >> representation". >> What is supposed to be used? I really can't imagine that the W3 >> committee responsible for the datatype definitions would really >> waste the time to define the canonical representation if they >> were not intended to be used. The schema WG carefully defined a set of lexical representations for each type, and in situations where more than one representaiton is legal for a given type, a canonical representation. So "3" and "003" are equally acceptable representations of the integer 3. >> Does anyone know to what level of strictness the various XML >> schema validation tools are enforcing the datatype value >> representations? To be compliant with the recommendation, all schema validation tools MUST accept all lexical representations of any given value. The schema recommendation mandates no situations in which the canonical form must be used. In practice, you will see the canonical form used quite a bit in situations where the value does not start out in XML. If a program in Java does: int i = 3; and you want to serialize that in XML, you'll find a lot of middleware that outputs "3" instead of "003". Validators MUST accept both forms. Particular applications of schema might mandate canonical form only, but schemas itself does not, and there is no single facet in the datatypes spec that can be used to enforce use of canonical values. ------------------------------------------------------------------ Noah Mendelsohn Voice: 1-617-693-4036 IBM Corporation Fax: 1-617-693-8676 One Rogers Street Cambridge, MA 02142 ------------------------------------------------------------------
Received on Tuesday, 12 February 2002 12:40:22 UTC