Good afternoon all, I have been mulling over some of the implications of the UCB proposed concept authority structure and I thought it might be useful to start a discussion thread to get input from other participants on these issues. Since many will be using this structure to create multiple authorities for work types, materials, process & techniques, style & periods, cultures, etc. it is important that we're sure the model we're building is the best possible approach. Some collections will want to enforce the use of the preferred term in the data record with variant terms being used for search only. Others will want to be able to use any term, regardless of its preferred or non-preferred status in a data record depending upon the context. The HAVRC collection uses the AAT (http://www.getty.edu/research/tools/vocabularies/aat/index.html) to control the terms used for work types. The AAT uses a faceted thesaurus model, where the “preferred” term and all the variant terms reside in a single data record with a single AAT id. For example, the term “miniatures (paintings)” (http://www.getty.edu/vow/AATFullDisplay?find=painting&logic=AND&note=visual&subjectid=300033936) has descriptors, alternate descriptors, and used for terms in 7 languages but all share the same AAT id 3000033936. Because the term is positioned within the hierarchy by AAT id, all 7 terms share a single position within the Visual Works hierarchy. If I understand correctly the way term completion works now in CollectionSpace, the authority term picklist depends on the enforcing the use of the preferred descriptor. If all of the alternate terms are recorded in a single CollectionSpace Concept Authority record in the repeating concept term group (modeling the AAT approach), the term flagged preferred will be the only term that shows in term list, thereby making it impossible to use an alternate descriptor (or even a preferred descriptor in a language other than English) in a data record. The implication here is that if a deployer wants to use the alternate descriptor “miniature (painting)” as a work type and “miniatures (paintings)” as a subject descriptor (or even as a work type for a complex work that was comprised of multiple miniatures) they would have to be entered as separate term records in order for them to both show up in the picklist. This also means both terms would have to be individually linked to broader and narrower contexts since they are considered distinct terms by CSpace. As one might imagine, it gets exponentially more complicated for collections who want to provide multilingual support. So this makes me wonder if there isn’t a way to build a fully functional clustered term thesaurus where any preferred or alternate term may be used in any data record without having to be a separate term, thereby simplifying the hierarchy. To be sure there would need to be some logic in the term completion that perhaps shows the preferred terms first, with the option to see and select the alternates if desired. Seems to me I saw a wireframe for something like this at one time but I don’t know if the idea went anywhere. The place authority that is now under construction here at UCB will also confront this same issue with historical place names. Any and all thoughts on these issues will be greatly appreciated. Thanks. Jan