RE: [plcs-dex] OWL and Reference data

Hi Mats

Some comments below

Regards
Rob

-------------------------------------------
Rob Bodington
Eurostep Limited
Web Page: http://www.eurostep.com http://www.share-a-space.com
Email: Rob.Bodington@eurostep.com
Phone: +44 (0)1454 270030
Mobile: +44 (0)7796 176 401

-----Original Message-----
From: mats.nilsson@fmv.se [mailto:mats.nilsson@fmv.se]
Sent: 30 May 2005 16:23
To: plcs-dex@lists.oasis-open.org
Subject: [plcs-dex] OWL and Reference data

Hi all,

I'd like to restart a discussion about reference data which if possible should result in a meeting session in Valencia (yes, I'm going there) dedicated to reference data.

[RBN>] I agree. Unfortunately, I will not be there. A high priority is to complete this initial set of PLCS reference data. This will involve taking all the modellers reference data from the individual modellers files, e.g.

dexlib/data/refdata/plcs_owl/plcs-rdl-lgm.owl

and moving it to:

dexlib/data/refdata/plcs_owl/plcs-proposed.owl
Sean, Leif, Mike and myself have drafted an initial set and updated this file. This needs to be reviewed and added to.

The main reason for bringing this topic up again is the need at FMV and the Swedish Defence to create/refresh a dictionary describing military and other common concepts. I've managed to make a number of the largest current projects interested in using OWL as the basic representation of the concepts/terms instead of creating an (e.g. Access based) application or using a simple (Word) document. The presentation to users would be in HTML created by XSL from the OWL-file (just like DEXlib or STEPmod). I believe you all can see some of the benefits this could lead to for us.

[RBN>] There is already XSL to present the OWL files to the users. See: dexlib/data/refdata/plcs_owl/sys/class_exp.xml

Do you have any suggestion as to how this presentation should be improved?
I don’t think that it deals with different languages for a start.

Therefore, I'd like to hear your comments on the following issues:
---------------------------------------
1. Label v.s. ID
I feel that the rdf:ID tag must be a unique identifier (within 10303?) which identifies the concept regardless of language. The rdfs:label tag should be used as the human interpreted way to distinguish between different concepts. This could be one way of handeling the 'homonym' problem, e.g. to separate "stone" (a primitive tool/resource for hammering…) from "stone" (a measurement for weight).

<owl:Class rdf:ID="plcs00001">
   <rdfs subClassOf rdf:resource="nnnn" />
   <rdfs:label xml:lang="en">identifier</rdfs:label>
   <rdfs:label xml:lang="se">identifierare</rdfs:label>
   ...
</owl:Class>

[RBN>] I agree – that is what ID is used for. Label should be used to provide alternative names, maybe in different laguages. To quote from:
http://www.w3.org/TR/2004/REC-owl-guide-20040210/#rdfs_label
The rdfs:label entry provides an optional human readable name for this class. Presentation tools can make use of it. The "lang" attribute provides support for multiple languages. A label is like a comment and contributes nothing to the logical interpretation of an ontology.

---------------------------------------
2. Minimum tag definition
We should agree on which of all possible tags and attributes should/must be populated in order for the reference data definition to be complete. This might not be a big task, but it might be an important checklist in the review process.

I guess these should be included;
<dcterms:created rdf:datatype="http://www.w3.org/2001/XMLSchema#string">2005-03-11</dcterms:created>
<dc:creator rdf:datatype="http://www.w3.org/2001/XMLSchema#string">Mats Nilsson, FMV, Sweden</dc:creator>
<dc:source rdf:datatype="http://www.w3.org/2001/XMLSchema#string">DEF-STAN 00-60</dc:source>

… and probably a few others describing version, status (proposed, defined, obsolete..) a.s.o.

[RBN>] I agree. When displaying the OWL classes, the XSL already flags an error if the following are not filled in:
dc:source dc:creator rdfs:comment
I agree that we need more. The online help should be updated to reflect this: dexlib/help/dex/rd-details.xml

---------------------------------------
3. Descriptions in different languages
I feel there should be a possibility to include a Swedish descriptive text (the rdfs:comment tag in OWL) to a concept defined e.g. in the PLCS core RDL, e.g. the concept 'Identifier'. This is important for the dictionary functionality described in the introduction above. How can we handle translations of the description? I believe there could also be a need for a Norweigan description, and a German, and…

There is no (as far as I've seen) implemented functionality in Protegé for descriptions in multiple languages. But it might be possible to use the rdfs:comment tag with an xml:lang attribute?! Is this allowed in OWL? Can there be multible rdfs:comment tags according to the schema?

In that case a possible solution is this;

<rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" xml:lang="en">
Specifies the number of weeks needed.
</rdfs:comment>
<rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" xml:lang="se">
Anger antal veckor som åtgår.
</rdfs:comment>

[RBN>] Yes, this is the correct approach, and yes you can already do this. Just add an rdfs:comment as another annotation property, selecting the language. Admittedly, Swedish does not feature as a language on the pull down menu- I am sure that this can be fixed.

---------------------------------------
4. 'about'/'ID' issue
Leif discovered in an OWL-file that the concept name sometimes appered in the rdf:ID tag and sometimes in the rdf:about tag. We need to figure ouf if this is a Protegé error or something else. Leif can probably explain this better.

[RBN>] OWL is represented as RDF and there is no guarantee as to how that RDF is output.
Sometimes you get: the class defined first

<owl:Class rdf:ID="Name"/>

Sometimes you get a reference to the class defined first, in which case you get

<owl:Class rdf:about="#Date_planned"/>

IN other words, the OWL tools do not necessarily produce RDF is a nicely ordered fashin. This makes writing any XSL against this a little more complex.
Is this what you meant?

---------------------------------------
5. Additional OWL tags
Can additional non OWL tags be used without messing up RDL functionality, e.g. for the versioning of descriptions of concepts and for other 'non standard' meta data regarding concepts. Or is perhaps everything already there?!

[RBN>] Yes, you can add other annotation properties. I think that we should define a standard set that we want and document what they are though, rather than allowing people to add their own.

Looking forward to hear your comments/suggestions/questions…, and to see you in Valencia!

Regards,
Mats
..............................................................................
Mats Nilsson – KC Logistik
FMV - Försvarets materielverk
Swedish Defense Materiel Administration
S-115 88 Stockholm, Sweden
tel: +46 8 782 47 24 mobile: +46 70 267 47 24
fax: +46 8 782 65 64 e-mail: mats.nilsson@fmv.se

plcs-dex message