- From: <noah_mendelsohn@us.ibm.com>
- Date: Fri, 25 Jan 2002 20:24:50 -0500
- To: xml-dist-app@w3.org
Me too...any examples showing particular programming languages should be in the primer, if at all. W3C Schema plays a distinguished role in SOAP, and should be discussed normatively. I suggest this not because it is technically superior to other schema languages, or because its use is mandated by SOAP (it mostly isn't) but because: a) schema is part of the W3C family of recommendations, so we have an obligation to show how protocols works with schema b) SOAP does formally mandate use of a few features of W3C XML Schema. Specifically, it adopts built-in simple types such as xsd:integer, and provides a SOAP-specific interpretation of xsi:type that applies even when schema validation is not performed. So, I suggest we need to show normatively which aspects of schema are mandatory in all SOAP implementations (you can't change the rules for what xsi:type="xsd:integer" means), and which are not. We need to clarify whether an element that says <A xsi:type="xsd:integer" >12XYZ34</A> is ever OK, and if not what the fault is and whether it MAY, MUST or SHOULD be triggered. Does the answer depend on whether we're using the "chapter 5" encodings? I think this must at least be discussed, but I'm not making a strong recommendation as to the answer...we probably can follow the precendent we just set for unresolved ID/IDREFs.) If we want to show how SOAP is used independent of W3C schema, we can show examples where no schema is presumed to exist, or we can present examples that presume the existence of a schema in some other hypothetical langauge, asserting certain constraints on the envelope. I would NOT use some particular known schema language such as schematron or Trex; I would just say "assume that a schema in some language constrains the <addresss> element to be non-empty", or some such. ------------------------------------------------------------------ Noah Mendelsohn Voice: 1-617-693-4036 IBM Corporation Fax: 1-617-693-8676 One Rogers Street Cambridge, MA 02142 ------------------------------------------------------------------
Received on Friday, 25 January 2002 20:32:07 UTC