[Om] SCOVO: Statistical Core Vocabulary

[Professor James Davenport]

On Wed, January 27, 2010 11:28 am, Christoph LANGE wrote:
> Dear James,
>
> 2010-01-27 09:43 Professor James Davenport <jhd at cs.bath.ac.uk>:
>> I  found this a useful reminder that we really ought to address the "RDF
>> and OpenMath" issue.
>
> I'm up to that; I have always been interested in exploring that
> connection.
> In my soon-to-be-finished thesis there will be a lot of elaboration on
> that.
Excellent.
> And, BTW, the way how I learnt about SCOVO is that a friend working on
> semantic web topics contacted me and asked me how they should reasonably
> express the mathematical relations between individual data points in RDF.
> E.g. if http://statistics.db/Germany#GDP2009 is described as having the
> numeric value V1, the unit €, some origin, further metadata, etc., if
> http://statistics.db/Germany#GDP2010 is described as having the numeric
> value V2, and if there is another data point
> http://statistics.db/Germany#GDPGrowth20092010 with numeric value V3 and
> "unit" percentage points, that the mathematical relation is V3 = V2/V1 *
> 100 - 100.
>
> Well, I said that it's certainly straightforward to write it down in
> Content MathML or OpenMath, using e.g. <csymbol
> definitionURL="http://statistics.db/Germany#GDP2009"/> or <OMS
> cdbase="http://statistics.db" cd="Germany" name="GDP2009"/> for the data
> points (the latter works here, but is problematic in general, as not all
> URIs from the RDF works have the form aaa / bbb # ccc).  It is less
> straightforward
> and maybe neither desirable nor reasonable to translate that to a pure RDF
That would be my opinion. If it WERE in RDF, how would he reason about it?
I would go for ANY semantic web construct with decent semantics, and
MathML-C/OpenMath is such.
> representation.  Massimo Marchiori suggested such a translation at MKM
> 2003,
> but to the best of my knowledge that never really been adopted.  RDF
> usually
> only supports binary predicates, plus certain cumbersome data structures
> that
> reasoners rarely support, but for adequately representing n-ary operators
> one
> would have to use those list- or array-like data structures.
>
> So in the end I suggested looking at one of the existing RDF-based unit
> ontologies; it might be the case that some of them have semiformal
> definitions
> of binary arithmetic operators allowing you to say (in RDF) e.g.
>
> S rdf:type :Sum.
> S :result X.
> S :firstArg Y.
> S :secondArg Z.
>
> Maybe that is actually the best way to go; and maybe it would be enough
> for us
> if the definition of that "Sum" "operator" were somehow linked to
> http://www.openmath.org/cd/arith1#plus.
Certainlyroom for a good debate here.

James Davenport
Recently: Visiting Full Professor, University of Waterloo
Now back as:
Lecturer on XX10190, CM30070, CM30078/50123, CM50209
Hebron & Medlock Professor of Information Technology, University of Bath
OpenMath Content Dictionary Editor and Programme Chair, OpenMath 2009
IMU Committee on Electronic Information and Communication