[foaf-dev] Mapping (tv and radio) programmes to their social media accounts

Ben Companjen bencompanjen at gmail.com
Wed May 30 18:25:50 EDT 2012

On 30 May 2012 22:17, Michael Smethurst <Michael.Smethurst at bbc.co.uk> wrote:
> probably garbled reply courtesy of webmail
>>> I've been planning to publish mappings from BBC programmes to their
>>> social
>>> media accounts on Facebook / Twitter / etc
>>> The easiest way would be to use foaf:Agent and foaf:account and say the
>>> foaf:Agent is same as po:Programme
>>> Can a programme be an agent? Would declaring same as break anything
>>> horribly? Or is there a neater way to do this?
>> Sounds plausible to me. "Agent" is always a slippery notion, and
>> because of that, it can save us sometimes from maybe needlessly rich
>> designs. I'm sure in some contexts you might want to model a
>> programme's associated teams in great detail, but for this case a
>> simple 'account' link sounds fine.
> cool, makes life simpler

Perhaps a bit too simple ;) - if you declare po:Programme to be the
same as (owl:sameAs) foaf:Agent, a reasoner would think that I,
because I consider myself a foaf:Person, am a Programme too. If you're
making it like Agent, make it a subclass, I'd say.

But why would you say any/every Programme is an Agent?
If because you want to use the FOAF properties to connect to social
media accounts/pages: why not create the same properties, but with
another domain? Doesn't SIOC have properties for this?
If because you really think programmes can be considered Agents: I
think Programmes don't "do stuff", as the FOAF spec and (IIRC) my AI
book define agents. IMHO Programmes are products of other agents,
mostly human agents.

Maybe this is too simple too, but since virtually every Programme is
created by a production team, why not connect the accounts to the
(perhaps unspecified) "<programme_name> Team" (and connect the team to
the programme with <dct:creator>)? They make the programme, possibly
interact on the social media accounts etc.


More information about the foaf-dev mailing list