Re: [ubl] Minutes from the Pacific Call (and call for urgent action from XML schema testers)

["Stephen Green" <stephen_green@seventhproject.co.uk>]

Greetings Folks

 

Tim in a report of the Pacific Call 2004-09-01/02 wrote 

1.3 Revised schemas for Unspecialized Data Types:
Following the Copenhagen plenary, a subgroup from the NDR team reviewed and recommended changes to the schema module used for Unspecialized Data Types.  This was in regard to the use of built-in XSD data types and the subsequent use of complexType or simpleTypes for these components. These are outlined (along with the complete set of new schemas) in the email found at http://lists.oasis-open.org/archives/ubl/200408/msg00090.html and paraphrased here...

Please find enclosed new schemas. I made changes to the
unspecializedDatatypes schema module. According to the NDR rules CTD9 and
CTD10 all unspecialised data types that are based on CCTs represented
through an simpleType now are simpleTypes and restrict the same xsd built-in
type like the appropriate CCT.

Changes were made to the following UDT:

DateTimeType
DateType
IndicatorType
NumericType
PercentType
RateType
TimeType
ValueType

As these changes have generated no comment or feedback we are concerned that no validation or impact study has been undertaken.  We understand that all other modules in the package listed in the above email are identical to those reviewed in Copenhagen. Therefore we would urgently request all members of the TC who normally do schema and instance testing to review this new schema module (attached in xsd and xsdrt form) along with the schemas accepted during the Copenhagen plenary and comment accordingly before Saturday 4th September.  (Action: All TC schema testers)

 

 

I haven't tested the Schemas with regard to XML/XSD syntax (other than validating with XML Spy 2004)

but I have performed the following test with 100% PASS of each part of the test.

 

For each document Schema in turn I generated (with a 'wizard' of XML Spy 2004 Enterprise on Windows XP) a sample instance

with all elements and all atributes of UBL and with generated, valid data *using the cd2 XSDRT Schemas*.

 

For each instance I performed a standard validation with XML Spy 2004 (please note this thros an error with XML Spy v5 which

those present in Copenhagen F2F with XSD knowledge all believed to be a 'bug' in the product and not the Schemas - due to

the use of restriction in the SDT Schema). RESULT: ALL VALID (there is sometimes a problem with the size of the OrderResponse

full sample instance but this merely required care to ensure the generator had finished generating before running the validation).

 

For each instance I changed the Schema location to point to the Schemas last sent by Gefeg (including the UDT Schema just

sent out by Tim) on 2004-08-24. I then revalidated. RESULT: ALL VALID

 

I attach a copy of the test instances.

 

This is not so much to test the XSD syntax so much as to ensure that no model change has occurred since the cd2 Schemas.

It also confirms correctness in quite a few other areas.

 

Many thanks to David / Gefeg for this Schema generation work.

 

All the best

 

Stephen Green

 

New Folder.zzz