Where the subset relationship breaks

[MURATA Makoto (FAMILY Given)]

Dear colleagues,

Since  master schemas are generated from S and T, it is now quite easy
to examine differences between S and T.  I tried to makesure if the 
subset relationship really holdes.  Apart from the namespace name
changes, I found some issues.

1) Namespace names as default values of the minVer attribute 
   in dml-diagram.xsd

Although this is not stated in prose of 29500:2008, the default value
of the minVer attribute in dml-diagram.xsd is a namespace name.  Since
the upcoming AM makes S and T use different namespaces, the default
values in S and those in T will be different.

Do we really need default values for @minVer? 

2) A required attribute @spid only in T

The attributeGroup AG_Ole in the transition pml.xsd contains

    <xsd:attribute name="spid" type="a:ST_ShapeID" use="required"/>

but this attribute is not present in the strict pml.xsd.  

This difference has existed since the publication of 29500.  Is this 
a defect or feature?

3) A required attribute @val only in T

The complex type CT_Charset in the transitional wml.xsd contains

<xsd:attribute name="val" type="ST_UcharHexNumber" use="required" />

but this attribute is not present in the strict wml.xsd.

This difference has existed since the publication of 29500.  Is this 
a defect or feature?

4) CT_OLeObject in pml.xsd

The transitional version makes @pic optional, while the strict version 
makes it mandatory.

Although this difference does not break subset relationship, is this
really intentional?  At least, this is not mentioned in 29500-4:2007.

This difference has existed since the publication of 29500.

5) CT_Font

In the strict version of sml.xsd, the complex type CT_Font follows the
def of CT_FontFamily; in the transitional version, it precedes it.  I
think that this is an editiorial mistake.


Cheers,
Makoto