Closed steads closed 7 years ago
From my work with the Getty, I disagree. Sometimes art researchers are very particular about the order in which things are listed, and this has semantics. Some examples of objects with several associations having the same Role. Maybe the order of these associations is significant.
Even for different roles, art researchers sometimes want to see them in a particular order.
This can be mapped as E13_Attribute_Assignment over the relation, using the extension property crmx:display_order
we've defined for CONA and JPGM
If there are semantics in the order then this needs to be made clear by using the appropriate CRM mechanism to show the semantics. So for instance the two "Sitters" examples above would be best modelled using CRMdig and the "possible attributions" by using CRMinf.
What properties from CRMdig and CRMinf express ordering?
The semantics are not actually about ordering. The ordering is being used to express the real semantics. So in the sitters example the semantics is about who is depicted adjacent to other subjects. The relative strength of possible attributions is about levels of believe.
Now you're ascribing meaning to the DisplayOrder field that may or may not be there. Maybe sitters are listed by importance, how would you know. I'm just saying if that field is present, we should emit it, because at least in some cases it has meaning.
In the CRM if we do not understand it we do not map it. This is axiomatic to the development of the CRM. Local solutions are fine. See our forthcoming CAA paper for guidelines.
Getty asked me to map it for CONA and JPGM, so they value it. As I said above, art researchers want the particular order.
DisplayOrder: This is a system implementation instruction and as such has no need for a mapping.