• this was dct:subject ... the label fell off. Note also that dct:subject can apparently point directly to Ron himself (I learned this recently)

The Topic Topic

Newer Older

Sketch of the RDF/OWL/SKOS technology landscape for describing topics...

derivadow, masaka, Roderic Page, and 6 other people added this photo to their favorites.

  1. masaka 63 months ago | reply

    nice summary. maybe missing 'subject' relation label on the bottom arrow (from HTML Regan to Concept Regan) ?

  2. PatHayes 63 months ago | reply

    OK, now I want to know what (the hell) a "conceptualization" is. Where should I start reading?

  3. leobard 63 months ago | reply

    in which namespace is "it" defined?

  4. danbri 63 months ago | reply

    @masaka - thanks, fixed in the original

    @pathayes - conceptualization is SKOS, which more or less is an RDF vocab for describing thesauri and similar things that are not quite directly understandable as class/property-based ontologies. See www.w3.org/TR/skos-reference/#concepts but please don't get too distracted by the "unit of thought" definition...

    @leobard ... i once proposed skos:it, but it was a bit premature. the relationship is not currently defined anywhere. If I persuade enough people it is useful, I'll put it into FOAF unless a better home appears...

  5. Ian Davis 57 months ago | reply

    I created a pair of properties open.vocab.org/terms/category and open.vocab.org/terms/isCategoryOf that might fill the role of "it" on your diagram

  6. danbri 51 months ago | reply

    see also wiki.foaf-project.org/w/term_focus and related discussion

  7. geckomarma 51 months ago | reply

    So if we have two URIs for Ron (himself) we connect them with ... ?

  8. PatHayes 51 months ago | reply

    sameAs, I would say, gecko.

  9. danbri 51 months ago | reply

    yes, owl:sameAs

    Is there a decent candidate spec for 'maybeSameAs' yet?

  10. PatHayes 51 months ago | reply

    NOt yet. We are trying to catch and preserve a few wild specimens in order to dissect them carefully, before trying the genetic engineering part. One insight (? maybe?) that is emerging is, chunks of published content are best thought of as having a touch of modality about them. So its not one huge WWRDFgraph, fully transparent.

  11. geckomarma 51 months ago | reply

    Ok, thanks re: sameAs. I'm asking because I'm currently revisiting our linked data implementation and want to link our "persons" to VIAF concepts and DBPedia ... things, among others, correctly.

    After seeing this I am thinking along the lines of www.flickr.com/photos/marma/4385930176/

  12. mlzeng 37 months ago | reply

    The pink area's legend may use the words "Conceptualizations (SKOS concepts)" or "Conceptualizations (@conceptScheme)" because SKOS itself does not conceptualize. --MZ

  13. a houghton 32 months ago | reply

    Why is "US Presidents" a subclass of foaf:Person? Seem like "US Presidents" should be a foaf:Group and "Ronald Regan" a foaf:Person who is a foaf:member of the foaf:Group "US Presidents".

  14. Michael Uschold 30 months ago | reply

    The class "US Presidents" should be called "US President" to have consistent number. Every US President is a Person, hence a subclass of Person.

  15. danbri 30 months ago | reply

    Michael - I agree; the plural was a mistake. I'm suprised how much discussion this quick scribbled diagram has generated, and entertained to be be having it on Flickr.

    Regarding relationship between Group and Person, there has been a bit of discussion lately about using OWL2 punning to allow a Group to also be seen as a subclass of Person. This might allow for the idiom where a Group has worldly characteristics, like a chairperson or logo or hometown; but also for groups that are defined by membership criteria a la OWL. Bit of a hack though...

keyboard shortcuts: previous photo next photo L view in light box F favorite < scroll film strip left > scroll film strip right ? show all shortcuts