Standard formal semantic vocabulary for academic course information?

[Jason Woodward]

Do you produce or consume semantic data representing your course directory? For example, Cornell's Food Science department is listed here in HTML markup.

We're looking to build and work with one, and would like to start with an established well-recognized vocabulary. Medium isn't important (XML, RDF, OWL, etc) for now because a pointer to one of those resources will likely lead to the rest (if they exist).

[Jason Woodward]

Cut-and-paste from a reply I made to several UWebD posts:


On Jul 11, 2009, at 10:23 AM, Brett Bieber wrote:

>Have you looked into XCRI-CAP? seems to be more of a UK thing, but
>there's some good work in there.
>http://www.xcri.org/

>We used that on our graduate bulletin/catalog -
>http://bulletin.unl.edu/CSCE/810?format=xml

>After trying to apply the same schema to our undergraduate bulletin we
>found that we were extending the XCRI format so much that it was
>easier to just build our own schema.

>http://creq.unl.edu/schema/courses.xsd
>http://creq.unl.edu/courses/public-view/all-courses/subject/AECN

Thanks for the XCRI pointer - that looks like a great starting point for what we're looking for. It looks like it could easily be extended to meet some of our needs like peoplesoft-specific course identifiers, cornell-specific venues, etc by creating types in our own namespaces. Seems pretty well thought out.

I took a quick look at your courses.xsd. What parts were missing from XCRI-CAP? I imagine you'd have to create your own local types, which I expect we'll have to do, but I'm wondering if there was a more fundamental structural problem in XCRI-CAP that I'm missing? I'd like to avoid putting too much effort into extending it without first understanding what roadblocks you ran into!

On Jul 11, 2009, at 11:38 PM, Jesse Rodgers wrote:


>A standard format would be great, as well as RDFa or a microformat.
>There's also some info here -
>http://microformats.org/wiki/course-catalog-examples
>also -
>http://microformats.org/wiki/hied-course-examples
>http://microformats.org/wiki/hied-course-formats
>http://microformats.org/wiki/hied-course-brainstorming

>A long standing project that I never really got into is the course information microformat. My masters research went a >different direction from the specific course information -- I looked at home page design patterns and content for general >standard formating instead.

>I would be happy to find some time for this again if others are willing to contribute. I see the 'hied' break suggests at least >one other person is interested.

>If adopted it would be hugely helpful... in Ontario that information has to be publicly available but at Waterloo is a bit >messy with no semantic mark up. You can parse it out as many student projects have demonstrated. But microformats >would make it so much easier.

>Jesse

I'm less interested in the microformat itself, but rather the vocabulary. I'm interested in producing a web service that would be consumed to produce microformat-ted HTML. The microformat seems like it might be a good vehicle for establishing / extending the existing vocabularies though, so I'm in.

Jason

[Lindsay Clark]

I haven’t heard of an established vocabulary for this kind of information, but would also be very interested.

This may be a situation for a VoCamp. At VoCamp, people work collaboratively on an ontology for a particular use. It may be worthwhile to have a University VoCamp.

-----------
Here’s what I know about what’s already going on:

Talis in the UK has been working on some interrelated ontologies for the academic industry. The main one is AIISO
http://vocab.org/aiiso/schema

I don’t think they have an ontology to represent courses at this point, but AIISO is being used in a project under development at Oxford (with JISC, the same group that worked on XCRI).

There is an ontology under development by Patrick Murray-John at University of Mary Washington. It is not established, but it really does offer a comprehensive model.
http://www.patrickgmj.net/project/university-ontology



Is anyone else interested in a University VoCamp?

-Lin
--
Lin Clark
Web Developer
University of Pittsburgh
[e-mail]
http://www.lin-clark.com

[Patrick Murray-John]

Lin et al.,
Thanks Lin for tweeting about this discussion! I'm definitely interested in a University VoCamp. I've been poking away at the university ontology you linked to a bit more, and will push up a revised and corrected version as quickly as I can.

I've looked at the XCRI schema. It might have the widest adoption of anything out there. I lean toward RDF, though, and don't know if there's an RDF implementation of it out there. From what I remember, it might not be too hard to pull together, and could probably reuse lots of AIISO.

The key thing, I think, in deciding which way to go is to think about what, exactly you want to model about the university and why. Part of the reason that there are these options is that each of them has a different worldview on university info. AIISO looks at both administrative and teaching related components, XCRI seems a bit more about just the catalog. The key thing that I'm trying to model is much more person-centric -- what people took what courses, what are their interests, what skills did they acquire, etc., along with the administrative and catalog info.

So each of the ontologies will be useful for providing the data to answer different types of questions. I'm curious to learn more about what kinds of questions people have in mind.

Thanks!
Patrick

[Jason Woodward]

I've attracted some interest from elsewhere within Cornell for working on this. The group produces http://vivo.cornell.edu which is, among other things, a browsable triple store of who's doing what at Cornell. They've got a lot of experience going between unstructured data, XML and RDF. I'm going to chat with them to see what their professional reaction is to AIISO and XCRI. They might be able to provide advice on the translations.

The reason I want to see XML come out of this is so it can be more easily consumable on the client side in JavaScript, versus RDF. I'd like to see both happen, though.

[Patrick Murray-John]

Interesting....I'll take a look at vivo.

On the javascript question, the rdf/json spec that Talis made has been pretty useful for all my needs. It's helped a lot that the rdf store I've been using, ARC, has a plugin that will serialize to that format. Not entirely sure if there are serializers to json for other stores.