N-relational entities (Was owl:Class and owl:Thing)

Hi Azamat,

<AA>
...for it is a real entity (like causality and space-time relations) rather
than an idea...
</AA>

We have indeed n-relational classes (which we call 'templates'), not just
n-relational predicates. These n-relational entities are Resources. See for
example the work of Natasha Noy and Alan Rector [1].

We can (and do) attach any kind of meta information to them (record creation
datetime, record creator (shown below), content meta data (e.g. "Approved
for Construction", a reference to a Change Order, etc), and/or, if
necessary, even a SPARQL script and/or rules.

An example of a template instance is:

    <tp:ST-3401 rdf:ID="ST-734651">
        <dm:recordCreator rdf:resource="#JOHN_DOE-20060125T1037Z"/>
        <tp:possessor rdf:resource="#FPO-347621-20060124T1645Z"/>
        <tp:propertyType
rdf:resource="http://www.15926.org/2006/02/rdl#MAWP"/>
        <tp:basePropertyType
rdf:resource="http://www.15926.org/2006/02/rdl#PRESSURE"/>
        <tp:numericalValue rdf:resource="#XSFL_100"/>
        <tp:unitOfMeasure
rdf:resource="http://www.15926.org/2006/02/rdl#BARG"/>
    </tp:ST-3401>

    <rdl:XML_SCHEMA_FLOAT rdf:ID="XSFL_100">
        <dm:content
rdf:datatype="http://www.w3.org/2001/XMLSchema#float">100</dm:content>
    </rdl:XML_SCHEMA_FLOAT>

where ST-3401 is the ID of the applicable template class in OWL, with the
label "quantified property of individual". We are building a vocabulary of
such template classes.

(Note the fact that we also turn Literals into Resources. That makes
translations easier (also represented by means of templates)).

Regards,
Hans

PS dm: = ISO 15926-2 data model in OWL, rdl: = ISO 15926-4 vocabulary of
reference data, tp: = ISO 15926-7 vocabulary of template classes

[1] http://www.w3.org/TR/swbp-n-aryRelations/

============================================================

-----Original Message-----
From: Azamat [mailto:abdoul@cytanet.com.cy] 
Sent: Monday, April 03, 2006 17:38
To: Hans Teijgeler
Cc: semantic-web@w3.org; ONTAC-WG General Discussion
Subject: Re: owl:Class and owl:Thing

''What exactly did you have in mind when you referred to "n-relational
ontology of things"? <HT/>

Hans:

Relations should be analyzed not only with respect to the number of terms
they connect and formal properties as cardinality, symmetry, transitivity,
reflexivity. This is all the subject of a formal relational logic. As old as
Russell statement that 'every proposition should be regarded as expressing a
relation between two and more things', like in the from R(x, y, z,...).
Above all, relationship should be considered with respect to the nature and
meanings of its components; for it is a real entity (like causality and
space-time relations) rather than an idea, although expressed by verbs and
propositions and abstract nouns and relational adjectives. Thus here comes
''n-relational ontology of entities''.

Regards,
Azamat Abdoullaev
http://www.eis.com.cy

-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.385 / Virus Database: 268.3.5/300 - Release Date: 03-Apr-06

Received on Tuesday, 4 April 2006 08:57:17 UTC