samvera-deprecated / sufia

[DEPRECATED] Sufia: a fully featured, flexible Samvera repository front-end.
http://sufia.io/
Other
111 stars 78 forks source link

RDF terms should be available in a public vocabulary #1327

Closed spr7b closed 7 years ago

spr7b commented 9 years ago

Did these terms ever get posted somewhere to make them part of a public vocabulary?

escowles commented 9 years ago

I wrote up a rough draft of an RDFS description of those predicates:

https://gist.github.com/escowles/93eea55a4341da25cdf5

Outstanding issues:

  1. Is http://scholarsphere.psu.edu/ns# the right namespace to use?
  2. Related: who can host these?
  3. Do the domains and ranges match how these are used in the new PCDM-ified Sufia?
awead commented 9 years ago

The Scholarsphere namespace seems a bit out of place now. Are they're any alternatives to use? @mjgiarlo, any issues for existing apps if we change these?

mjgiarlo commented 9 years ago

@awead Existing apps will have to migrate data to the new predicate URI, yep.

awead commented 9 years ago

@mjgiarlo maybe something to pull the plug on for Sufia 7.0 with PCDM, since we'll be redefining a lot predicates anyway.

mjgiarlo commented 9 years ago

ooooh, I like that idea. I'll ticket that.

jcoyne commented 9 years ago

These are currently in curation_concerns, so if we're going to change them we should do it ASAP.

mjgiarlo commented 9 years ago

Good point re: relativePath and importUrl being in CC. I'll look at @escowles's gist today and come back with a proposal.

mjgiarlo commented 9 years ago
  1. No, that's not the right namespace.
  2. How about opaquenamespace.org? Or is there a more appropriate/stable place? (Hey, how stable would you consider opaquenamespace.org, @tpendragon @kestlund @no-reply? A penny for your thoughts about where to host a wee Hydra-related vocab, preferably somewhere that isn't going to go away in the next few years?)
  3. About domains and ranges, are pcdm:Files a valid domain given the wonky way that we express metadata about files in F4? If so, OK. The range of relativePath should be a string given the current impl.

Thanks for getting this started, @escowles!

ruebot commented 9 years ago

Greetings!

@jlhardes contacted @acoburn and I, and wanted us to check out this issue in our roles as co-facilitators for the Technical Metadata Working Group. @acoburn and I had a call today and talked about it, and are more than happy to share our feedback now, or if you'd like, we can pull the original group together and consult them as well.

cheers! :smile:

awead commented 9 years ago

@jlhardes @acoburn @ruebot It just so happens there's an agenda item for this on today's Hydra tech call. https://wiki.duraspace.org/display/hydra/Hydra+Tech+Call+2015-10-28 Care to join in?

ruebot commented 9 years ago

I'd love too, but I have a schedule conflict at noon :cry:

jlhardes commented 9 years ago

I can be on the call today. @ruebot @acoburn can you go ahead and share your feedback? I think that will help move this forward.

acoburn commented 9 years ago

I will also be on the call

ruebot commented 9 years ago

Thanks @acoburn!

jlhardes commented 9 years ago

Hi! @acoburn and @ruebot should verify this, but from what I have in my notes from the Hydra Tech Call on 2015-10-28, the following were predicate options to use instead of creating new terms:

premis: http://www.loc.gov/premis/rdf/v1 marcrelator: http://id.loc.gov/vocabulary/relators ebucore: http://www.ebu.ch/metadata/ontologies/ebucore/ebucore

ss:arkivoChecksum = premis:hasMessageDigest or a checksum property from LC (not sure which) ss:relativePath = premis:hasContentLocationValue or a property from ebucore (ebucore:locator?) ss:importURL = premis:historyInfo but this might not work ss:proxyDespoitor = marcrelator:dpt (depositor) ss:onBehalfOf = marcrelator:cre (creator)

hackartisan commented 9 years ago

Some of these did not seem like great fits as I recall, especially the marcrelators for proxyDepositor and onBehalfOf. My sense from the conversation was that more investigation / discussion would be needed. @cam156 seemed to have the best sense of how these terms were actually being used.

ruebot commented 9 years ago

@jlhardes that sounds about right.

mjgiarlo commented 8 years ago

Given no clear path to resolution at the moment, I am moving this to the backlog so that it's not a blocker to 7.0.0.

mjgiarlo commented 8 years ago

This issue was moved to projecthydra-labs/curation_concerns#561

mjgiarlo commented 8 years ago

I copied this issue to CC because the five terms originally specified are split between Sufia and CC. Tweaking the description in each to clarify which terms live where.

mjgiarlo commented 7 years ago

I believe the Hydra Metadata folks continue to work on this. Once we have a path forward, we should write up a new ticket for Hyrax.