supportDesc: schematron constraint

[mjhawkins]

There are several rules to debate:

1) When msDesc doesn't have a msPart child, physDesc/objectDesc/supportDesc's material attribute must be have content (ie. parchment, paper, etc)

2) When msDesc has an msPart child, either that msDesc or one of its ancestors must has a supportDesc with a material attribute with a value

3) the material attribute is recommended and is a semi-closed list. Doe we keep the changes to this attribute in the base ODD? If so, are we happy with the values, do we want to add more?

• chart
• mixed
• other
• palm
• papyrus
• perg
• unknown

[mjhawkins]

Leave supportDesc's material attribute test in GS unchanged.

The contents of the attribute will still need to be sorted out.

[mjhawkins]

Remove the check on the ancestor and require the presence of the material attribute on supportDesc all the time?