Vocabularies in ISO 15926 (we can use SBVR)
November 27, 2009
We have much of discussions here in Russia about translation of ISO 15926-2 into Russian.
But we have equal discussions about “best terms” when we have terminology difference in variety of industries.
ISO 15926 is a way to concept mapping. It is not work with homonyms and synonyms. It lack vocabulary layer. It is not so with Gellish that have a means to work with synonyms and homonyms.
I think that we can adopt approach of relatively fresh work of Object Management Group that issued SBVR (semantic business vocabulary and rules) standard. It is somehow similar to ISO 15926. It define:
— very brief upper ontology (like ISO 15926-2)
— very brief definitions to small amount of enterprise domain concepts — taxonomy (like ISO 15926-4)
— examples of presentation in 3 different language formalisms including FOL
— terminology ontology
— business rules concepts (that have only 15% of all text of standard despite mentioned in title!)
I think we can have ISO 15926-12 (?) to declare scheme how to deal with vocabularies (terminology, concept “labels”) to represent ontology of ISO 15926 items, prototemplates, templates, OIMs etc.. We can borrow terminology ontology of SBVR not as Yet Another Schema but as ISO 15926 common infrastructure part that deal with representation of concepts and meanings in different national languages and industry dialects.
Info about SBVR is here:
Text of OMG SBVR 1.0 standard: http://www.omg.org/cgi-bin/doc?formal/08-01-02.pdf
Text of tutorial for ISO TC37 that deal with terminology: http://home.btconnect.com/BusinessSemantic/How_SBVR_Adds_Knowledge_Richness_to_ISO_TC_37_Terminology_Standards.pps
Overview (in Gellish-like tables of concepts and relations) — http://www.kdmanalytics.com/sbvr/sbvr_overview.php (with Exel spreadsheet of SBVR elements).
SBVR is also suggested to semantic integration of SOA and Process in the same manner that ISO 15926 community suggest ISO 15926 (see presentation http://metadata-standards.org/Document-library/Documents-by-number/WG2-N1301-N1350/WG2_N1330_BabaPipranORM2009iPaper11_4.pdf from slide 59) — but with stressing terminology/linguistic and representational (including non-text representations) issues.
E-mail: 
Anatoly Levenchuk's Runglish Blog 
February 1, 2010 at 11:35
Hi Anatoly, you said …
“ISO 15926 is a way to concept mapping. It is not work with homonyms and synonyms. It lack vocabulary layer.”
This is not actually true … these vocabulary layers are in the RDL content … there are homonyms and synonyms (using identfiers in context, different names for the same thing in any context, and same name for different things in different context). Simplified templates (like Gellish, and including Gellish itself) can be used to construct this content – lierally called homonym and synonym. So far there are few in any homonyms, but that is the “convention” of working in one language to develop the content in current tools, but there are many synomym examples – they are just not called homonyms and synonyms.
February 1, 2010 at 12:59
Well, I know about simplest work on “labels” level like in Gellish-molecular-syntax. But I think about all terminology domain constructs like in SBVR and SKOS and this is not only concept of synonyms and homonims. I am understand that this can be modeled (or meta-modeled) in ISO 15926 language and even in (proto?)templates, but I think it is needed be somehow standardized (at least understand what linguistic-not-ontology-oriented standards we can take as a model for this layer: SBVR, SKOS or something else. Thus, we provided mr.David Leal with SBVR and he tell us some thoughts (11th of December 2010):
“1) I had not looked at the SBVR before, but I can see the relevance. Two key
aspects of the relationship between SBVR and ISO 15926 to be worked on are:
– use of 4D ontology: This is orthogonal to SBVR as far as I can see, so
that ISO 15026 could add a 4D approach to the SBVR.
– merging of type and extension: This is a very useful simplification for
much (perhaps all) engineering data. However some philosophers do not like it
because it relies upon members of classes in “hypothetical worlds”. It could
be useful to document how the SBVR is simplified by this approach.
2) Another interesting aspect is the relationship between SBVR and RDF (with
RDF named graphs) and OWL. Solving this problem may be very similar to the
relationship between ISO 15926 and RDF (with RDF names graphs) and OWL. A
solution for SBVR will probably work for ISO 15926 and vice versa.
3) SKOS seems to be a small, but useful, subset of the SBVR implemented
using RDF.
4) Gellish is a useful vocabulary which would work well within the SBVR
environment.”
He asked any ideas on how to take this forward but we now have no ISO 15926 training to suggest such an ideas. Yet we know that vocabulary/linguistic layer is very important and keep pushing this issue to people that knowledgeable enough to solve it.