OpenAtlas INDIGO Vocabs Meeting 2022-04-05¶

Updated information in the course of the meeting is in color. Every participant is invited to add and adapt.

Participants¶

• Alexander Watzinger
• Bernhard Koschiček-Krombholz
• Klaus Illmayer
• Massimiliano Carloni

This meeting is about importing thesaurus from ACDH-CH Vocabs into OpenAtlas (#1663). We invited our Vocabs expert Klaus to help clear up open questions.

Vocabs frontend¶

Is there a possibility/tool to enter and manage data directly in Vocabs (a SKOSMOS representation of a thesaurus)?

• ACDH-CH tool vocabseditor for simple solutions, but currently unmaintained.
• Vocbench, an external tool but very complex and with some unsolved issues currently.

Massimiliano will take a look at the vocabseditor but most likely we keep the current workflow (INDIGO providing the data and Massimiliano importing it to Vocabs).

Identifier¶

At the moment in the INDIGO project identifier like 0080 are used, are there better alternatives?
Klaus suggested using camel case notation of the label in the URI when stable, like e.g https://vocabs.dariah.eu/tadirah/en/page/audioAnnotation.

Versioning¶

Versioning is possible in theory but there are no good practical solution which could be used as examples currently. E.g. former versions could be archived in Arche and on Vocabs only the newest version is available. Problem with this solution would be for ones who still want to use an older version because they would have to update their systems to link to archived instance in Arche.
Additional suggestion from Klaus: don't use version numbers in URLs. Older version should be dumped with changelog and the new entries are linked to the old entries with close match.

Structure¶

Although collections/groups are good for a nice representation, they could be problematic for data entry/usage because they are not a concept and just containers. E.g. we have hierarchically groups, like Graffito Component but it shouldn't be used for linking in OpenAtlas. Klaus answered an in between question from Alex about broader/broadMatch: broadMatch is used for external references, broader for internal.

How to proceed¶

Before we can continue work on the OpenAtlas/Vocabs import script following issues would have to be solved by INDIGO/Vocabs:
1) Adapt identifier, e.g. if using an integer or names, either way we should change the current 0081 notation. Even if using integer it should be like this: 81.
2) We can't use the group hierarchy because it's "just cosmetically", we can only link to concepts in OpenAtlas, e.g.

• Graffito Component (top group is ok, it is used as label in forms but cannot be linked to)
  • Concept Idea (not ok, it is not a concept but would be link-able if imported into OpenAtlas)
    • Crossing Out (ok because it is a concept)
    • Greeting (ok because it is a concept)

There would be multiple approaches to solve this, e.g. dropping the 2nd level (Concept Idea) in OpenAtlas and display all 3rd level entries like a flat list. Another option would be to use concepts as 2nd level entries.