LRMI in Dublin Core

17 views
Skip to first unread message

Hugh Paterson III

unread,
May 12, 2021, 4:11:12 AM5/12/21
to lr...@googlegroups.com
Greetings,

I am working on a mapping between LRMI and how the concepts contained therein are expressed in, or could be expressed in Dublin Core (e.g., OAI-PMH). The issue I am approaching is: "How can a client use LRMI as an application profile, but not be bound to schema.org encoding"? Or more specifically, "How can an institution participate in LRMI as an application profile without moving beyond the OAI-PMH encoding scheme"? — often something their current technology and schemas already support.

1. Does anyone have experience with this already?

The pathway that walked down is to look at Library of Congress Subject headings for educational resources. OAI-PMH usually uses Dublin Core and Dublin Core Terms (unless an extension is explicitly declared — which one of the communities I participate in does do, but extending these schemas is more difficult than using solutions which are already provided). Since Dublin Core allows the explicit declaration of Library of Congress subject headings (either as literals or non-literals), it seems that subject headings related to educational resources could then be mapped back to LRMI concepts which are not explicitly available via the Dublin Core elements or terms. One use case of this would be for the identification of "resource type" as "textbook". I specifically found the Library of Congress Genre/Form Terms, which has a list of 52 educational types—including textbook.

In contrast to normal LCSH which describe the *content*, Genre/Form Terms describe the nature of the work (this is likely why they were split out from the LCSH term list in 2011). Even though they were split out I am uncertain if they are managed as a subsection of LCSH or as their own thing. That is, are they "valid" to use within Dubin Core's LCSH refinement (Vocabulary Encoding Scheme)  or even the LCC refinement of subjects?

Does anyone on this list have experience in this area? or know where I can ask questions about the usage of Genre/Form Terms within Vocabulary Encoding Scheme applied to subjects of Dublin Core? 

In one sense non-literals are non-literals so who cares what non-literals are put into a subject within OAI_DC. But in another sense, the value of non-literals is increased as the number of users of those non-literals increases, so the utility increases as the recognition of the non-literal increases. — Use the popular non-literal and more machines will find the advertised "thing". Add the associated literal and more people will find the thing. Use both and both machines and people will find it.

all the best,
- Hugh


Phil Barker

unread,
May 12, 2021, 7:00:19 AM5/12/21
to lr...@googlegroups.com, Thomas Baker

Hello Hugh,

One mapping between LRMI/schema.org and DCMI, I recall Tom Baker had a project, I think with an intern student, looking at this. I don't recall what the outcome was. I've copied this message to him so he may contact you.

My own thoughts are that such a mapping is superficially easy (dc:title is related to schema:name) but gets harder the more you get into the depths of schema and the more strict you want your mapping to be (e.g. dealing with how schema.org allows Literals for pretty much everything). It would easier to do for specific applications of schema.org than for the general case.

On the other points you mention, yes we did look at Library of Congress Genre/Form Terms when looking at creating a concept scheme for learning resource types. CEDS also has usable concepts for learning resource type, e.g. https://ceds.ed.gov/element/000928/#Textbook . It was a while back, but certainly something like

{
 "@context": {
    "sdo": "http://schema.org/",
    "gen": "http://id.loc.gov/authorities/genreForms/",
    "@base": "http://example.org/resources"
  },
 "@id": "#IntroToQuant" ,
 "sdo:learningResourceType": {"@id": "gen:gf2014026191"}
}

would be perfectly valid.

As far as I know/recall OAI-PMH requires metadata in XML, and the DC in OAI_DC is simple Dublin Core, so you might have to map text values in instance data to LOC IRIs.

Hope this helps, Phil

--
You received this message because you are subscribed to the Google Groups "Learning Resource Metadata Initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lrmi+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/lrmi/CAE%3D3Ky8iZt58CjW7eOvuz-QTwjCAbgRkzaucioU%2BV7mVLiB3Jg%40mail.gmail.com.
--

Phil Barker. http://people.pjjk.net/phil
CETIS LLP: a cooperative consultancy for innovation in education technology.
PJJK Limited: technology to enhance learning; information systems for education.

CETIS is a co-operative limited liability partnership, registered in England number OC399090
PJJK Limited is registered in Scotland as a private limited company, number SC569282.

Reply all
Reply to author
Forward
0 new messages