- From: Martin Gudgin <marting@develop.com>
- Date: Thu, 15 Nov 2001 08:47:09 -0500
- To: "MacAndrew, Tim" <tmacandrew@NetSilicon.com>, "Henrik Frystyk Nielsen" <henrikn@microsoft.com>, <Noah_Mendelsohn@lotus.com>, <xml-dist-app@w3.org>
- Cc: <xschema@us.ibm.com>, "Jones, Matthew" <MJones@NetSilicon.com>
----- Original Message ----- From: "MacAndrew, Tim" <tmacandrew@NetSilicon.com> To: "Henrik Frystyk Nielsen" <henrikn@microsoft.com>; <Noah_Mendelsohn@lotus.com>; <xml-dist-app@w3.org> Cc: <xschema@us.ibm.com>; "Jones, Matthew" <MJones@NetSilicon.com> Sent: Wednesday, November 14, 2001 3:59 PM Subject: RE: Formalism in SOAP spec > If I could add some analysis ... > > I sent the XML Schema for SOAP 1.2 > http://www.w3.org/2001/09/soap-encoding through > the "Schema Quality Checker" from IBM. The following > errors were generated: > > Initializing Schema Quality Checker. Please wait ... > SchemaQualityChecker has been initialized > PROCESSING C:\Work\SOAP\soap_enc_orig.xsd time : > 2001-11-14T12:44:02 ... > C:\Work\SOAP\soap_enc_orig.xsd (file 1 of 1) now being read ... > ERROR > file = file:C:/Work/SOAP/soap_enc_orig.xsd line 224 column 55 > SEVERITY: 1 > ERROR TYPE: 2 > MESSAGE > The global simpleType > http://www.w3.org/2001/09/soap-encoding:NOTATION > is not found. > > ------------------------------------------------------------------ > ERROR > file = file:C:/Work/SOAP/soap_enc_orig.xsd line 231 column 20 > SEVERITY: 1 > ERROR TYPE: 2 > MESSAGE > complexType http://www.w3.org/2001/09/soap-encoding:NOTATION > requires the following facets : > enumeration > > ------------------------------------------------------------------ > > I believe there are two error conditions that are > causing this: > > 1) An <attribute> (specifically the > "NOTATION" attribute) cannot have a > 'type' that is a <complexType>. > Perhaps the NOTATION <complexType> should've > been declared as a <simpleType> ... All of the soap-encoding types basically add id and href attributes. soap-encoding always uses elements. > or > perhaps the <attribute> NOTATION was > really meant to be an <element>. Sorry, that's an error ( and I think it's my fault ). It is my recollection that the Working Group agreed a while ago that the NOTATION stuff should be removed from the encoding schema. It was on the editor's TODO list but for some reason never got done. > > 2) Use of the NOTATION (built-in Schema > type) requires that (one of?) the facets: > length > minLength > maxLength > pattern > enumeration > whiteSpace > be used. This was one reason that we decided to ditch the notation type from the schema. The other is, as you point out above, it only works for attributes and soap doesn't use attributes. > > Please note that I had to add the XML prolog: > <?xml version = '1.0' encoding = 'UTF-8'?> > to the Schema document ... which I believe is required. No, XML Declaration is not required if encoding is UTF-8 or UTF-16 > > Also, if I could suggest that naming of an > <attribute> (e.g. "NOTATION") that is the same as > a <complexType> that is also a "built-in" Schema type > is confusing. Although it is "legal" Schema (an > exercise in namespacing?), it is difficult to read. Schema has six non-overlapping symbol spaces. So there is no confusion on the part of a software although I agree it may sometimes be confusing for wetware. Regards Martin Gudgin DevelopMentor
Received on Thursday, 15 November 2001 08:47:59 UTC