[expertfinder-dev] Re: ExpertFinder Lunch Meeting/Expertise Vocab

[Tom Heath]

Hi Paola,

Thanks for the message. I've merged your two mails so I can reply as one :)

On 19/06/07, paola.dimaio at gmail.com <paola.dimaio at gmail.com> wrote:
> Tom
>
>  Great stuff, thanks for starting this
>  In order to comment, we need to be able to read your work first!
>  How shall we best view your definitions?
>
>  how can I best view your rdf file? (does not open in rss reader)

The links should open fine in a browser such as Firefox, although you
may need to view-source to see the RDF code. Unfortunately I don't
think an RSS reader would be generic enough to parse this.

>  also, is there a human readable interface of this somewhere?

Unfortunately not right now, as the work is only at an early stage.
I'm current interfacing with it just via a text editor ;)

>  shall we have some iterations on a wiki page or similar

Yes, possibly, although I think it would be interesting to gather a
range of opinions on this list first. I have a very specific use case
for this, and the Hoonoh vocab is based on the results of a previous
study which I need to adhere to in pursuing this research further. So,
in that context there isn't much scope for me to change the vocab that
*I* use (apart from modelling decisions perhaps). If this same vocab
(or just a few classes or properties) can be useful to the wider
expertfinder community then that's great, I'd be really pleased. If on
the other hand the community decides it needs to adopt a different
approach or vocab that's fine too, of course :)

>  I may have missed some of the progress, so apologies for knowledge gaps#
>  In my paper I made a proposal that we add a property to foaf called
> 'expertise'
>  has this been considered?

As a community my personal opinion is that we should follow the SIOC
approach, and integrate as closely as possible with FOAF, whilst not
trying to actually add additional properties to the spec. I think
trying to add new properties to FOAF itself may delay us
unnecessarily.

>  the value would then be the different kinds of expertise etc

Do you mean "different forms of expertise", or "expertise with regard
to different topics"?

>  There is a difference beween expertise and experience, in the sense that
> expertise arises from experience, whereby experience is a tangible thing
> that can be related to
>  precise time occurrencies, while expertise is more an abstract quality and
> the possibly deriving from the sum of all experiences

Yes, I like the point that experience can be related to specific
events in time. At a more general level I'm personally a little
hesitant in assuming anything about the relationship between
experience and expertise, as expertise can take many forms, can be
very subjective, and in some cases may fail to emerge irrespective of
the amount of experience gained.

For example, a family member of mine may be my personal "expert" on
the subject of building and home improvements, but without having any
qualifications, just lots of knowledge and experience. Another person
may not trust him on this topic to the same extent as I do (and it's
useful to be able to model/capture this, which the Hoonoh vocab will
allow by using named graphs I guess, although we don't currently
generate subjective "expertise views"; this may be covered by the grey
area between Expertise and Affinity relationships??).

In contrast my doctor is an expert in general healthcare, by virtue of
a qualification and associated experience. Her qualification may
result in many more people trusting her as an expert by default than
would trust my family member about home improvements. Lastly a friend
of mine travels to Brussel frequently for business, but generally
flies there and back the same day, spending most of the day in
meetings. So, he has plenty of experience of being in Brussels, but as
far as I know he is no kind of Brussels expert.

So, I think we probably agree that expertise is a very multi-faceted
concept ;) The challenge is to create a vocabulary that's simple, but
sufficiently flexible to allow these subtleties to be represented.

Cheers,

Tom.

>  so i would start from defining the property name, and after discuss the
> rest of the issues?  makes sense?
>
>
>  cheers
>  Paola Di Maio
>
>
>
> <hoonoh:ExpertiseRelationship
> rdf:about="http://some-uri-for-this-expertise-relationship/">
>      <hoonoh:from rdf:resource="http://example.com/people/jack/" />
>      <hoonoh:toTopic rdf:resource="http://example.com/topics/beer/" />
>      <hoonoh:value>0.532</hoonoh:value>
>  </hoonoh:ExpertiseRelationship>
>
>  <hoonoh:ExperienceRelationship
> rdf:about="http://some-uri-for-this-experience-relationship/">
>      <hoonoh:from rdf:resource="http://example.com/people/jack/" />
>      <hoonoh:toTopic rdf:resource="http://example.com/topics/beer/" />
>      <hoonoh:value>0.895</hoonoh:value>
>  </hoonoh:ExperienceRelationship>
>
>  <hoonoh:AffinityRelationship
> rdf:about="http://some-uri-for-this-affinity-relationship/">
>      <hoonoh:from rdf:resource="http://example.com/people/jack/" />
>      <hoonoh:toPerson rdf:resource="http://example.com/people/jill/" />
>      <hoonoh:value>0.20</hoonoh:value>
>  </hoonoh:AffinityRelationship>
>
>  </rdf:RDF>