Re: SW Graphical Notation

Dear all,

in light of this discussion about ORM as a visualisation - as well as a "fact-oriented" modelling paradigm for ontologies I refer to all the work that has been done in the DOGMA project at VUB STARLab over the past 14 years. This approach has been succesful in many domains. In particular it was one of the USPs for the spinoff Collibra I cofounded 5 years ago.

Business domain experts are particularly attracted by fact-orientated modelling, the visual diagrams as well as the more verbose rule formatting (see SBVR), and they do not need very expressive rules in most cases. This trade-off between expressivity and usability dramatically reduces the perspicuity bottleneck of ontology engineering, and indirectly increases the reusability of "ontology bases".

Some early papers to reflect on where ORM meets :
- Robert Meersman: The Use of Lexicons and Other Computer-Linguistic Tools in Semantics, Design and Cooperation of Database Systems. CODAS 1999: 1-
- Robert Meersman: Semantic Ontology Tools in IS Design. ISMIS 1999: 30-45
- Peter Spyns, Robert Meersman, Mustafa Jarrar: Data Modelling versus Ontology Engineering. SIGMOD Record 31(4): 12-17 (2002)
- Pieter De Leenheer, Aldo de Moor, Robert Meersman: Context Dependency Management in Ontology Engineering: A Formal Approach. J. Data Semantics 8: 26-56 (2007)

Collibra's Business semantics Glossary (as part of its Data Governance product) is reported:
-on the site: www.collibra.com<http://www.collibra.com>
-Lior Limonad, Pieter De Leenheer, Mark H. Linehan, Rick Hull, Roman Vaculín: Ontology of Dynamic Entities. ER 2012: 345-358

Happy to continue this discussion anytime,
Cheers,


-- Dr. Pieter De Leenheer

Assistant Professor of Computer Science - VU University Amsterdam
http://www.pieterdeleenheer.be<http://www.pieterdeleenheer.be/>



On 28 Jun 2013, at 15:08, Enrico Franconi <franconi@inf.unibz.it<mailto:franconi@inf.unibz.it>>
 wrote:


On 28 Jun 2013, at 14:42, Eric Prud'hommeaux <eric@w3.org<mailto:eric@w3.org>> wrote:

ER, UML, ORM statements are inherently graphical in nature and the graphics actually tells you in an intuitive way their meaning.

I've seen a couple which seemed to help people grok schemas. There's an early OWL (or maybe DAML+OIL) doc that I stumble across from time to time which represented partitions as radially divide elipses (sort of like pie charts).

Sure. This is in the spirit to represent graphically the aspects for which the visual representation actually helps. By the way, the graphical representation of a partition in ORM reminds also the "sliced filled pie" <http://www.orm.net/images/subtype_pic.png>.

I tried to put triples and set constraints in <http://www.w3.org/TR/hcls-kb/#triplemodel>. These may not be super-intuitive for the person who has never considered intersections and unions, but them hollowed or solid diamonds and circles aren't apparent until one has learned the UML visual language.

It is not obvious to me that a representation of classes and instances with explicit Venn diagrams actually helps, since it may become quite tangled with large non-tree taxonomies, but it is worth exploring I guess.

cheers
--e.