Re: reconciliation of disparate models - Jon

Jon:

    The problem is that some system has to 'make sense'
    of that resource and that _:blank1 resource ID has
    to be interpreted as a non-ID, so it's existence as a
    resource is ephemeral. Again, if it's worth describing
    (or necessary) then it's worth giving it a URI, even one
    your system coins, and making it a real thing.

             dcterms:publisher http://mydomain/publisher/12345 ;
         http://mydomain/publisher/12345 a foaf:Organization
           foaf:name "Harper Collins";

    ...is much better, and this lets other systems understand
    and create valid relationships (like owl:sameAs) between
    your 'internal' resource and their understanding of
    the same resource in their and other systems. How will
    your data work in the aggregate of the open web? What's
    'correct' in almost every case when there's a question of
    how to publish RDF is found in trying to figure out how
    the data is likely to be understood and used by consuming
    systems, made harder of course by the fact that you can
    only guess.

-- 
Tom Baker <tbaker@tbaker.de>

Received on Sunday, 13 March 2011 23:43:47 UTC