Skip to content

Instantly share code, notes, and snippets.

@mark-dce
Last active March 18, 2019 16:00
Show Gist options
  • Save mark-dce/550cfe5b2f63afe4ca06cd2f4adca04c to your computer and use it in GitHub Desktop.
Save mark-dce/550cfe5b2f63afe4ca06cd2f4adca04c to your computer and use it in GitHub Desktop.
Potential CSV model for collections and child pages
Object Type Title Item ARK Collection ARK Parent ARK Sequence Number
Collection Miscellaneous Manuscripts ark:9999/1df68c5
Work Home Book on Sanitation ark:9999/2fa8801 ark:9999/1df68c5
Child Work Front Cover ark:9999/8c500dc ark:9999/2fa8801 1
Child Work Inside Front ark:9999/6899e32 ark:9999/2fa8801 2
@mark-dce
Copy link
Author

mark-dce commented Mar 17, 2019

Main Ideas:

  • Add an Object Type column to distinguish between Collections, Works, and sub-Works (pages for now, but potentially other complex structures later)
  • Everything has a unique ARK including page-level images
  • Collections don't generally have parents, but this format leaves room for nested collections if needed
  • Collections and Works aren't ordered, so sequence number is optional and ignored
  • Child works are ordered within the work they're attached to, so have an explicit order

Potential Issues:

  • Should we omit sequence on child works and just use the row order - upon discussion, sequence number should be required to help diagnose potential display and ingest issues.
  • Pages (child works) end up having a small amount of required metadata - is it reasonable to assume you just copy the parent metadata in the absence of more specific info for the children (i.e. keywords are required for works right now, should we require page-level keywords?)

@bess
Copy link

bess commented Mar 18, 2019

I'd like to see an explicit sequence number... if we're relying on row ordering it's too easy to lose a row and not realize it.

Is it safe to assume we'll register an ARK for each page object? Is that the current practice?

@mark-dce
Copy link
Author

Agreed about row order - requiring an explicit sequence number makes sense.

All of the examples we have seen (e.g. http://digital2.library.ucla.edu/viewItem.do?ark=21198/zz000sq41q) have page-level ARKs. UCLA has indicated this will also be the case going forward, so we'll use this as our working assumption until proven wrong.

Based on discussion, we're dividing out the child-level parent vs. the work's collection parent.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment