use of cei:idno and atom:id could be clearer

[GVogeler]

There is a bunch of questions around the identifiers in MOM-CA and this issue is to try to collect them:

1. atom:id in private collections is an uuid with no semantic value. Some users request a meaningfull URL derived from cei:idno ("slug name", "permalink").

2. cei:text/cei:idno is used very inconsistently. Some archives/collections copy the collection information into the cei:idno, some use an abstract value (e.g. derived by date), some use a local numbering system representing chronlogical order.

3. The main purpose of cei:text/cei:idno is to identifiy a document in its current context. So each of the above solutions is correct. But consider a hooked-in-charter: It brings its local cei:idno with it into the "hosting" collection, while atom:id is then a new uuid. In this context the cei:idno can be non-unique, which is not desirable.

4. cei:idno can be used as a local identifier (cei:text/cei:idno) and as part of the archival identifier in the cei:witness sections (e.g. cei:witnessOrig/cei:archIdentifier/cei:idno). Archives very often don't give the second information as the main cei:text/cei:idno is identical during import of a full fonds.

Anything else?

[larkvi]

When converting this to TEI P5, I think the important dictinction is "idno should be used for labels which identify an object or concept in a formal cataloguing system such as a database or an RDF store, or in a distributed system such as the World Wide Web." Since cei:idno is not unique to a charter (3.) and not used consistently (2.), one wonders whether our use of idno really counts as formal.

(4.) points to a distinct use of idno that exists in TEI, as well. One, as the header element uniquely identifying a resource, the other as the shelfmark in <msidentifier>.