Hi Creighton,
Keen eye! I've taken an initial look at this, and I think the issue arises from the fact that at present, AtoM only has 1 field that doubles as both 1.8B18 and 1.8B20 - you will notice if you click into the AtoM Associated materials field that the tooltip contains quotes from both rules. I don't know why this decision was made (template design predates me!), but it's what we're working with at present.
As you found in issue #
5211, this meant that we were dropping <separatedmaterial> elements on import. Since the EAD tag library associates both <separatedmaterial> and <relatedmaterial> with ISAD 3.5.3 (Related units of description), the easiest short term fix for 5211 was simply to make sure it imported. We are also crosswalking the RAD Allied materials field to the ISAD Related units of description, which is likely why we defaulted to using the EAD <relatedmaterials> element on export for this field. The ISAD standard used Related units for both "descriptions in the same repository or elsewhere" while RAD has these 3 different elements in the standard, but only the one field in AtoM.
It is also worth noting that at present, the RAD "Related materials" autocomplete field (which you are not using) is not actually added to the EAD XML at all.
So... I can file a ticket for this, but what is the desired outcome?
To me, it seems that the ideal outcome would probably be:
- Add 2 new fields to the RAD template for 1.8B20 and 1.8B20a
- Rename the "Related materials" autocomplete field to "Related descriptions" (as it is in the ISAD template)
- Add separate EAD mappings for the 2 new RAD fields, and update the import / export, search index, CSV templates, and XSLTs
That quickly becomes a project that is beyond the scope of what Artefactual can do without community sponsorship, however.
Barring that, what would you consider to be the best mapping for this field? Should it no longer crosswalk to the ISAD Related units field? Should we keep the crosswalk, but have the ISAD template use <relatedmaterials> on export, but RAD use <separatedmaterial>? Something else? I welcome your input! If we hit upon a good solution, I would be happy to file a bug report.
Cheers,