...
- We're most interested in integrations that are generalizable and shareable. Strictly speaking, we're not interested in purely local integrations.
- A plugin can be an integration, shuttling data between different sources can be an integration, and changes to the data model in various ways are integrations.
- We're looking for minimal manual intervention to qualify as an integration. That is, integrations do not have a manual step through, even if that step through is automated. For example, a script that, at a given interval, checks to see what Resources have been updated in ArchivesSpace, and automates the export of MARC records to a watch folder that an ILS checks periodically and uses to import and overlay existing MARC records, while involving separate systems, is not an integration.
Tiers:
To be determined, but the lowest might be a plugin that sits on top of the data model without changing it and the highest would be like Yale's container management work that requires changes to the data model and migration of data already in the system.
...