Re: Modelling Resumes in RDF(S) - work in progress

On 2002-06-05 2:34, "ext Uldis Bojars" <uldis.bojars@gmx.net> wrote:

> 
> Hi!
> 
> I am working on master thesis about modelling of Resume information
> in Semantic Web using RDF.
> 
> And I have a couple of questions for WWW-RDF-Interest. Any feedback is
> welcome.
> 
> _1_ About describing Person information.
> 
> For describing personal (contact) information I find vCard RDF schema
> most complete with properties that need to be described. [ FOAF is also
> good for describing person's info, but I guess there is no sense to duplicate
> both vCard and FOAF entries for a person. ]
> 
> You can see very rough draft of structure of CV data to be included in
> the model at: http://blackeye.vsaa.lv/~davidson/scheme.gif
> (it's handwritten; letter M next to properties that appear in CV entry
> forms in monster.com; underlined "M" indicates mandatory properties)
> 
> The properties needed are:
> name, address(es), phone(s)/fax/pager, e-mail(s), UID(s), birth date and
> place, webpage, sex, nationality, citizenship.
> 
> vCard handles most of them (as literals), but what ontologies to use for:
> 
> # Describing sex - literals "M"/"F"?
>                  - WordNet - Man/Woman?

Choosing URIs as values allows one to localize the interface
using alternate labels in other languages. Literal values would
be less internationally oriented.

E.g. 

   <cv:Sex rdf:about="voc:blackeye.vsaa.lv/cv/sex/male">
      <rdfs:label xml:lang="en">Male</rdfs:label>
      <rdfs:label xml:lang="fi">Mies</rdfs:label>
      ...
   </cv:Sex>

C.f. [1]

> # Describing citizenship - country codes from ISO 3166-1 ?

Sounds like a good choice, though TGN might also suffice (see below)

> # Describing nationality - is there an ontology for nationalities?
> 
> There are no properties for birth place in vCard, but that can be
> extended with a couple of literal properties.

For birth place values, have a look at TGN:

http://www.getty.edu/research/tools/vocabulary/tgn/

and see [1] for how these can be encoded as URIs for minimal
graph redundancy (see examples)

> 
> _2_  About linking Person with a CV
> 
> I defined classes Person (subclass of wn:Person), which will describe
> the person and CV (subclass of wn:Curriculum_Vitae), which will contain
> CV information entries.
> 
> What would be the best way to link there two classes?
> I think it is reasonable to have person as a separate class.
> At the same time CV is the main subject what I am speaking about.
> 
> So - I would add a property aboutPerson to a CV class and then point
> this property to a person class. Or is it sufficient to point to
> person information in <cv:CV about="person_info">?

This would not give you the results you need. It essentially
defines the person to be of class CV rather than saying its
the CV about the person.

Rather, you want the aboutPerson property, as in

   <cv:CV rdf:about="{URI denoting CV}">
      <cv:aboutPerson rdf:resource="{URI denoting person}"/>
      ...
   </cv:CV>
 
> _3_ Structure of CV information
> 
> Structure of CV information as I imagine it is:
> Class CV and class CV_Entry which has subclasses for each type of CV
> entries (work hist, education, ...). Class CV has properties
> hasWorkHistory, hasEducation, ..., which point to appropriate CV
> entries.
>
> Is this structure OK for describing CV information?
> Are there better suggestions?

Looks fine to me, though I didn't see how you plan to
relate the actual content with the CV_Entries.

Cheers,

Patrick


[1] http://ietf.org/internet-drafts/draft-pstickler-voc-01.txt

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Wednesday, 5 June 2002 04:47:40 UTC