Schema for machine-readable institutional metadata

[thatbudakguy]

As discussed in geoblacklight/geoblacklight#987, it would benefit institutions who harvest and index OGM records to have a reliable way to show the user contact information corresponding to the institution whose record they are viewing. In Earthworks, this is currently just a hard-coded list of names and emails.

This ticket is a proposal: what if this information were to be offered as part of a single machine-readable file stored at the root of each institution's OGM repository? The file could use a common interchange format (e.g. JSON) and contain similar data to what is currently suggested for inclusion in the README: contact information, the schema version used for the metadata, the style of persistent URL/link used (if any), etc. The schema of this record would be defined and shared on the opengeometadata.org website.

Then, GeoBlacklight and/or GeoCombine could programmatically read this file and use it to display the relevant information while the user is browsing records in any GeoBlacklight instance (which would solve geoblacklight/geoblacklight#987). It could also be used to validate compliance with geoblacklight schemas: if the repository claims compliance with one standard but some or all of its records do not match that standard, this could be determined automatically.

[karenmajewicz]

This seems like a great idea - we have long talked about establishing some kind of template for basic repository documentation. (This issue is closed, but we didn't really accomplish it: OpenGeoMetadata/shared-repository#17). Having the template be machine readable is even better.

I searched around a bit to see if there was already a schema that would serve this purpose, but didn't really find anything. However, these standard GitHub files might work somehow?

Citation (CFF) files:

CODEOWNERS

[thatbudakguy]

CODEOWNERS is a really interesting idea, for a couple reasons:

• It's easy to list multiple owners
• GitHub shows a special UI hint that indicates who "owns" a file when you view it in GitHub
• Code owners are automatically requested for review when someone opens a pull request that modifies code that they own
• You can have separate owners for different files (i.e. someone could "own" FGDC metadata but not geoblacklight metadata)

One possible hangup is that you might want the "contact person" for an institution to not necessarily be the "code owner" of the metadata on GitHub...but I guess that goes back to what the point of "contact"ing someone is. If it's usually going to be something along the lines of "this record seems wrong/needs to be taken down", maybe the person who can do that on GitHub actually is the best contact person (rather than a non-GitHub using librarian or staff member)?

[karenmajewicz]

Interesting idea - revisit when we work on improving documentation for OGM repos