W3C home > Mailing lists > Public > public-owl-wg@w3.org > October 2008

Annotations and SKOS

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Tue, 7 Oct 2008 11:40:07 -0400
To: W3C OWL Working Group <public-owl-wg@w3.org>
Message-Id: <936FCAA4-92A8-458A-8309-6D7DAA970E47@gmail.com>
Following is a thread that lead into a discussion of using OWL 2  
annotations.
http://www.w3.org/mid/000701c92202$bc335710$349a0530$@tice@k-int.com

Below I include the discussion that I had with one of the posters,  
who makes comments on our description of annotations as well as their  
suitability to her task. I thought it might be a useful input to  
discussing SKOS/Annotations, and hope that we could respond to her  
most recent comment, if even to ask for clarification.

-Alan


Forwarded conversation

From: Sini, Margherita (KCEW) <Margherita.Sini@fao.org>
Date: Mon, Sep 29, 2008 at 4:55 AM
To: De Smedt Johan <Johan.DeSmedt@wkb.be>, Rob Tice <rob.tice@k- 
int.com>, public-esw-thes@w3.org



If a concept is reused across different concept schemes my proposal  
would be
to have different namespace (and so different URI) for it and then  
define
them with the different labels as needed, and then use mappings.

E.g.
       uri c_in_a = http://myschemeA#c123  and has preferred label a
       uri c_in_b = http://myschemeA#c123 <http://myschemeA#c123>    
and has
preferred label b

http://myschemeA#c123 <http://myschemea/#c123>  exactMatch
http://myschemeA#c123 <http://myschemea/#c123>

Hope this helps.


        -----Or iginal Message-----
        From: public-esw-thes-request@w3.org on behalf of De Smedt Johan
        Sent: Mon 9/29/2008 09:48
        To: Rob Tice; public-esw-thes@w3.org
        Cc:
        Subject: RE: further SKOS question

        http://www.k-int.com <http://www.k-int.com/>


----------
From: Sini, Margherita (KCEW) <Margherita.Sini@fao.org>
Date: Mon, Sep 29, 2008 at 4:56 AM
To: De Smedt Johan <Johan.DeSmedt@wkb.be>, Rob Tice <rob.tice@k- 
int.com>, public-esw-thes@w3.org



(sorry corrected mistake)
               uri c_in_a = http://myschemeA#c123 <http://myschemea/ 
#c123>
               uri c_in_b = http://myschemeB#c123 <http:// 
myschemeB#c123>
http://myschemeB#c123 <http://myschemeB#c123>


----------
From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Tue, Sep 30, 2008 at 12:46 AM
To: "Sini, Margherita (KCEW)" <Margherita.Sini@fao.org>
Cc: De Smedt Johan <Johan.DeSmedt@wkb.be>, Rob Tice <rob.tice@k- 
int.com>, public-esw-thes@w3.org


There is a proposal for how to annotate any axiom, including  
annotations, in OWL 2. I have a professional interest in getting an  
opinion on whether that would serve the purpose you need. The  
documentation is available at: http://www.w3.org/2007/OWL/wiki/ 
Syntax#Annotations

In principle you should be able to use this to say anything about an  
annotation such as a label, including which concept scheme it is a  
preferred label for. Of course this would need some adjustment of the  
SKOS schema.

Regards,

Alan Ruttenberg
co-chair OWL working group ;-)


----------
From: Sini, Margherita (KCEW) <Margherita.Sini@fao.org>
Date: Tue, Sep 30, 2008 at 1:08 AM
To: Alan Ruttenberg <alanruttenberg@gmail.com>


Thanks Alan,

I will surely have a look to OWL 2... I need to catch-up with it....

I also recently posted an issue on public-owl-dev@w3.org... I do not  
know if
you have seen it ("Modelling in OWL relationships depending by multiple
variables") but I was wondering  if i would need OWL 2 to solve that  
issue.

Regards
Margherita

<http://myschemea/#c123>  <http://myschemea/#c123>
<http://myschemeb/#c123>  <http://myschemeB#c123 <http://myschemeb/ 
#c123> >
                http://myschemeB#c123 <http://myschemeb/#c123>
<http://myschemeB#c123 <http://myschemeb/#c123> >


----------
From: Sini, Margherita (KCEW) <Margherita.Sini@fao.org>
Date: Fri, Oct 3, 2008 at 3:40 AM
To: Alan Ruttenberg <alanruttenberg@gmail.com>


Hi Alan,

I had a look to the section you pointed me for OWL2: 10 Annotations
(but i wish to find the time to read all the chapters)

You mention that these annotations "does not affect the logical  
meaning of
the ontology" or "should be ignored during document parsing". But in  
para
10.2: Is not clear to me what means  "annotations affect structural
equivalence between axioms"... why in the example you provided the 2  
axioms
are NOT structurally equivalent? They just differe by a comment,  
which should
not have a menaing in the logic of the onto...

Anyway, what i need is that if i have a concept, i can create  
relationships
between labels (if defined as rdfs:label)... for example i want to  
specify
that "FAO" is an acronym of "Food and Agric. Org."..., where they are  
both
rdfs:label of the same concept... but in OWL 1 this is not possible  
(as i
know...).

With OWL2, with your annotations I am not sure i can achieve my  
objective...
Its true I can specify the other problem about identifying that my  
label is
scientific or a common name, but can i wish also to do relationships  
between
rdfs:labels...

Hope this helps.

<http://myschemea/#c123>  <http://myschemea/#c123>
<http://myschemeb/#c123>  <http://myschemeB#c123 <http://myschemeb/ 
#c123> >
                http://myschemeB#c123 <http://myschemeb/#c123>
<http://myschemeB#c123 <http://myschemeb/#c123> >


----------
From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Fri, Oct 3, 2008 at 11:58 AM
To: "Sini, Margherita (KCEW)" <Margherita.Sini@fao.org>


Hi Margherita,

Would you mind if I send this on to the documentation authors of the
spec? I think it would be helpful feedback.
Some comments in line.

On Fri, Oct 3, 2008 at 3:40 AM, Sini, Margherita (KCEW)
Structural equivalence is about the format of documents. Saying two
things are not structurally equivalent just means they don't have the
same content, not that they have the same entailments. So, for example
two document with

A: SomeValuesFrom(hasChild person)
B: MinCardinality(1, hasChild, person)

Express the same logical statement but are not *structurally*  
equivalent.
Does the below do the trick? It annotates the assertion that FAO is
the label of fao with the property acronymFor value Food and Agric.
Org."
In other words that "FAO" is the acronym is a comment on the labeling  
axiom.

Class(fao label(Annotation(acronymFor  "Food and Agric. Org.") "FAO" ))


-Alan


----------
From: Sini, Margherita (KCEW) <Margherita.Sini@fao.org>
Date: Sun, Oct 5, 2008 at 9:23 AM
To: Alan Ruttenberg <alanruttenberg@gmail.com>


Sure, feel free to send this email to other people or list. Maybe in  
this
case i may formulate other needs...
for example make annotations on properties (because we would like to  
mark
properties with author name, creation date, etc.) but this i could  
see has
been taken in consideration by OWL 2.

Your suggestion on the annotation may be ok for easy cases, but if we  
have
too many labels, each with its own language, then it may be difficult to
differenciate acronyms from spelling variants, from synonyms, from
tranlations... I.e.  we would like to make these kinds of relationships
between labels:

'relatedTerm'
'hasAcronym'
'isAcronymOf'
'translation'
'hasSynonym'
'hasNarrowerSynonym'
'hasBroaderSynonym'
'hasLocalName'
'isLocalNameOf'
'hasTradeName'
'isTradeNameOf'
'hasSpellingVariant'
'scientificTaxonomicNameOf'
'hasScientificTaxonomicName'
'isChemicalFormulaOf'
'hasChemicalFormula'
'isSymbolFor'
'hasSymbol'
'isAbbreviationOf'
'hasAbbreviation'
'hasSingular'
'hasPlural'
'hasDialectalVariant'
'isDialectalVariantOf'

Also in case of 'hasDialectalVariant' we may need to add a specific  
dialect
name... so maybe we would need more than annotations...

Hope this helps

<http://myschemeb/#c123 <http://myschemeb/#c123> > >
Received on Tuesday, 7 October 2008 15:40:54 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 7 October 2008 15:40:55 GMT