To outline the issue:
Take a child label which is 'related to' a parent label, and apply them both to an image. This results in every child of the parent being recorded in the XMP-ics data for that image.
For example, in my catalogue the label Queen Victoria public house is 'related to' the generic label pub. If I apply both labels to an image then the XMP-ics data for that image includes details of every specific pub 'related to' pub in my catalogue.
At the moment pub is only in 20 relationships in my catalog, but in time it's likely to reach hundreds - all of them cross-referencing each other in the XMP-ics data. And the same for each other set of relationships. That's going to grow into a very large amount of redundant data (in the technical sense of data redundancy).
Should consideration be given to (optionally?) handling this in a different way? Perhaps:
- Only recording the single relationship from 'child' to 'parent' in the XMP-ics data (my favorite provided the 'direction' of the relationship can be determined)?
- Optionally excluding relationships from XMP-ics?
- Enabling relationship data to be separately exported, in a similar way to the way that the catalog label hierarchy can be exported to a file?