Unresolved questions about the differences between Open Metadata Schema vs. GSA CSV input

[harrisj]

I was reading over the Project Open Metadata schema again, and I noted a few things that might be issues in the CSV format used by the GSA's team for specifying new entries for the data.json file. I wanted to outline them here so we can follow up with them on it.

• [ ] The modified field should be an interval if a dataset is continually updated. Do they use accrualPeriodicity instead of the Modified column of the spreadsheet if the dataset's periodicity is not irregular? This makes sense for datasets updated every minute, but what about a dataset that is updated once a month?
• [ ] The schema specifies that contactPoint is just This is a container for two fields that together make up the contact information for the dataset. The vCard format assumes this is a person, but does it really need to be? I just want to confirm this.
• [ ] The schema supports a landingPage field which is human-friendly landing page for the data that is not necessarily a data dictionary. Do we need to support that?
• [ ] It looks like the describedBy and describedByType fields are meant to be used for human-readable descriptions when at the top-level and machine-readable when within distribution records.
• [ ] The schema definition allows users to specify one or more references for documentation that is not a data dictionary. Do we have a need to use this anywhere?

It also looks like we have some issues with how we have represented APIs in our sample CSV. Will follow up with a related issue

[harrisj]

We will follow up on these after we've gone though a successful submission through the GSA process

[harrisj]

It looks like other agencies at least allow non-personal email addresses as the POC. See http://catalog.data.gov/dataset/medicaid-enrollment-by-state for instance