W3C home > Mailing lists > Public > public-bpmlod@w3.org > October 2015

Re: Questions about TBX to RDF handling

From: Felix Sasaki <fsasaki@w3.org>
Date: Tue, 27 Oct 2015 17:59:12 +0900
Cc: "public-bpmlod@w3.org" <public-bpmlod@w3.org>
Message-Id: <7955AD95-7C11-40C7-93C3-AC449EDEE63C@w3.org>
To: Peter Svanberg <Peter.Svanberg@tnc.se>
Hi Peter,

> Am 27.10.2015 um 01:24 schrieb Peter Svanberg <Peter.Svanberg@tnc.se>:
> 
> Hello Felix!
>  
> Thanks for your comments. I’ve read a lot about RDF, but where do you find such info, that rdf:li and rdf:Seq is not widely used …?


I cannot give you concrete statistics on this, maybe others in this group can help with this.

>  
> The URI is not the problem, I have (but didn’t show):
>  
> :Bas4711  a             skos:Concept ;
>         tbx:definition  "kommersiell <hi target=\"Bas4712\">verksamhet</hi> som drivs under ordnade former"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral <http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>> .
>  
> <http://example.com/f%C3%B6retag-sv <http://example.com/f%C3%B6retag-sv>>
>         a                      ontolex:LexicalEntry ;
>         ontolex:canonicalForm  <http://example.com/f%C3%B6retag-sv#CanonicalForm <http://example.com/f%C3%B6retag-sv#CanonicalForm>> ;
>         ontolex:language       <http://www.lexvo.org/page/iso639-3/swe <http://www.lexvo.org/page/iso639-3/swe>> ;
>         ontolex:sense          <http://example.com/f%C3%B6retag-sv#Sense <http://example.com/f%C3%B6retag-sv#Sense>> . 
>  
> <http://example.com/f%C3%B6retag-sv#Sense <http://example.com/f%C3%B6retag-sv#Sense>>
>         ontolex:reference  :Bas4711 .
>  
>  
> <http://example.com/f%C3%B6retag-sv#CanonicalForm <http://example.com/f%C3%B6retag-sv#CanonicalForm>>
>         ontolex:writtenRep  "företag"@sv .
>  
> :Bas4712  a             skos:Concept ;
>  
> :verksamhet-sv  a              ontolex:LexicalEntry ;
>         ontolex:canonicalForm  <http://example.com/verksamhet-sv#CanonicalForm <http://example.com/verksamhet-sv#CanonicalForm>> ;
>         ontolex:language       <http://www.lexvo.org/page/iso639-3/swe <http://www.lexvo.org/page/iso639-3/swe>> ;
>         ontolex:sense          <http://example.com/verksamhet-sv#Sense <http://example.com/verksamhet-sv#Sense>> .
>  
> <http://example.com/verksamhet-sv#CanonicalForm <http://example.com/verksamhet-sv#CanonicalForm>>
>         ontolex:writtenRep  "verksamhet"@sv .
>  
> <http://example.com/aktivitet-sv#Sense <http://example.com/aktivitet-sv#Sense>>
>         ontolex:reference  :Bas4716 .
>  
> :
> :
>  
>  
> So, a cross reference in text like this is never made into triplets? You require the “consuming end” to parse the XML?

I would rather propose to generate triples and not use XMLLiteral for the items that you want to represent explicitly. 

>  
> I just realized that a problem you get if you maketbx:definition into a XMLLiteral instead of a String(as it currently is in the OWL) is how to specify the language of the definition. In the proof-of-concept example European Migration Network glossary, http://lemon-model.net/lemon#definition <http://lemon-model.net/lemon#definition> is used, placed in a pseudo-node, to allow for further description of the definition. Could a pseudo-node be used also for tbx:definition? I.e.:
>  
> :Bas4711  a             skos:Concept ;
>         tbx:definition           Bas4711#def
>  
>  
> :Bas4711#def          a          tbx:definition ;
> tbx:value             "kommersiell <hi target=\"Bas4712\">verksamhet</hi> som drivs under ordnade former"^^<http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral <http://www.w3.org/1999/02/22-rdf-syntax-ns#XMLLiteral>> ;
>            ontolex:language       <http://www.lexvo.org/page/iso639-3/swe <http://www.lexvo.org/page/iso639-3/swe>> .

That is again with XMLLiteral, so maybe not the best approach.

>  
>  
> Maybe I should also mention that so far, usage of explicit/clickable cross references in definitions is not widely used in our environment. Non-obivous terms which are used in definitions are assumed to refer to concepts in the same terminology source (a certain set of term entries) and if the term is not unique within that source, a homonym number is appended. Important concepts in the definition is sometimes put into separate “see also” fields.
>  
>  
> swedish term (preferred): 
> företag 
> definition: 
> kommersiell verksamhet (2) som drivs under ordnade former
> see also: 
> verksamhet (2) <http://www.rikstermbanken.se/rtb/visaTermpost.html?id=205091>
>  
> But that is partly for historical reasons. When we convert to RDF, where links are so important, we would like to be able to make real links in definitions. Will that put an unusual burden on the “consuming end”?


I don’t think so - making the links explicit would be great.

Best,

Felix

>  
> /Peter Svanberg
>  
> Från: Felix Sasaki [mailto:fsasaki@w3.org <mailto:fsasaki@w3.org>] 
> Skickat: den 24 oktober 2015 08:44
> Till: Peter Svanberg
> Kopia: public-bpmlod@w3.org <mailto:public-bpmlod@w3.org>
> Ämne: Re: Questions about TBX to RDF handling
>  
> Hello Peter,
>  
> other people who are more knowledgable in the TBX to RDF handing may come back to you, bellow some general feedback.
>  
> Am 22.10.2015 um 16:42 schrieb Peter Svanberg <Peter.Svanberg@tnc.se <mailto:Peter.Svanberg@tnc.se>>:
>  
> Returning again to this issue, specifically the cross reference. I got the comment that having XML code in the definition entry text is not so RDF-ish – how should it be made into triplets?
>  
> After looking more on RDF handling, I suppose you could make it into a ref:seq, something like:
>  
> <rdf:Description rdf:about="http://example.com/Bas4711 <http://example.com/Bas4711>">
>     <tbx:definition rdf:parseType=”?”>
>         <rdf:Seq>
>             <rdf:li>kommersiell</rdf:li>
>             <rdf:li ref:resource=”#Bas4712”>verksamhet</rdf:li>
>       
>     <rdf:li>som drivs under ordnade former</rdf:li>
> </rdf:Seq>
>     </tbx:definition>
> </rdf:Description>
>  
>  
> The rdf:li and rdf:Seq constructs are not used widely, I would try to avoid them. 
> 
> 
> which then should be concatenated to display the definition. Or how do you handle ”marked up” text in RDF?
>  
> The general way is parseType literal, what is in your mail below. But this does not give you a URI for kommersiell or verksamhet, and I think this is what you want. 
>  
> <http://example.com/Bas4711 <http://example.com/Bas4711>> tbd:definition _bs393.
>  
> _bs393 rdfs:label „kommersiell“.
> _bs393 rdfs:label „verksamhet“.
>  
> Would that make sense for you?
>  
> Best,
>  
> Felix
> 
> 
> And need these things be described in your recommendations?
>  
> /Peter
>  
> 28 sep. 2015 kl. 17:01 skrev Peter Svanberg <Peter.Svanberg@tnc.se <mailto:Peter.Svanberg@tnc.se>>:
>  
> 2.       We would like to use cross references in some fields, manly definition. TBX allows that through a hi element in the value, for example (text in Swedish, sorry …):
> 
>       <termEntry>
>         <langSet xml:lang="sv" id="Bas4711">
>           <descrip type="definition">kommersiell <hi target="Bas4712">verksamhet</hi> som drivs under ordnade former</descrip>
>           <tig>
>             <term>företag</term>
>           </tig>
>         </langSet>
>       </termEntry>
>  
>  
> :
> :
>  
>   <rdf:Description rdf:about="http://example.com/Bas4711 <http://example.com/Bas4711>">
>    <tbx:definition rdf:parseType="Literal">kommersiell
>                <hi target="Bas4712">verksamhet</hi> som drivs under
>        ordnade former</tbx:definition>
>     <rdf:type rdf:resource="http://www.w3.org/2004/02/skos/core#Concept <http://www.w3.org/2004/02/skos/core#Concept>"/>
>   </rdf:Description>
Received on Tuesday, 27 October 2015 08:59:30 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 27 October 2015 08:59:30 UTC