| Because you asked --- The primary purpose of a semantic model is to facilitate knowledge exchange and choice making in a gaggle of humans in hopes of morphing the gaggle into a system. A key usage is to inform the development of an executable ontology, e.g., application software, for automation of information flow and decision. Another key purpose is to provide a basis for objective assessment of enterprise situation (aka evidence-based management).
To my earlier point, until a prospect values purpose and usage it is largely futile to gain their awareness, appreciation and acceptance of ontology development or adoption of a pre-canned ontology. In fact, if they say they want an ontology it is likely that they are 'silver bullet' or "latest fad" connoisseurs and will not become successful customers.
In the case I described five distinct outcomes occurred. They reviewed their data base schema in light of the semantic model. They were amazed to find approximately 50,000 business rules then astonished to discover that approximately half counteracted the other half. They became aware of how inadequate their current IT was and the probable cost and schedule of creating a new one. They declined the 'opportunity" to buy the pre-canned insurance enterprise information model they were being offered because it didn't seem to fit their business. They were already involved in corporate strategic maneuvers which cannot be described here. They decided to spin off their IT department to serve other insurance and related businesses based on their new-found abilities in modeling and object technology.
The moral? Ontology is not a hammer. Ontology is a mirror.
Jack Ring On Mar 1, 2011, at 8:51 AM, Michael F Uschold wrote: Thanks for that example. I second what John Sowa said, the key is to know what is really going on in the business, and to translate that into a suitable structured notation.
Jack, you mentioned the idea of fit-for-purpose ontology, but I saw no mention of the intended purpose of the 'ontology' (aka stunning semantic model) that you built was for?
In my experience, most people who say they want an ontology don't really know why, other than it seems like a good idea. There are many possible uses, but you can't build a fit for purpose ontology w/o a purpose in mind!
It would be a great addition to your story, how the client actualy made USE of the ontology/semantic model you delivered.
Michael On Mon, Feb 28, 2011 at 7:23 PM, Jack Ring <jring7@xxxxxxxxx> wrote:
In the mid-1990's Tom Love's object technology practice was busy introducing Smalltalk-based software to the marketplace.
Two projects had been enjoined with major insurance companies. Both projects directly engaged Smalltalk class designers with insurance-domain specialists. Progress was slow on both projects even though both were staffed with 90 percentile object technology practitioners. After six months a lot of refactoring had happened but only a sparse class library had been achieved.
A third customer signed up for which I became responsible. Instead of the 8-15 Smalltalkers as on the first two projects there was only one senior and on fresh-out available for this third insurance project. Time to innovate. Instead of engaging in class library design Doug McDavid proposed that he discover what insurance people really talk about. After processing 200 or so documents ranging from Annual Reports to underwriter risk analyses he produced a stunning semantic map. Two more Smalltalkers showed up and in two months time we had a class library designed and reviewed that was "better" than either of the other two projects in their ninth month.
The cost comparison was approximately $X for us and $11X for each of them.
I think this experience highlighted two factors.
1) Because business activities entail lots of knowledge exchange and choice making the coherency of the persons' respective mental models is key to adequate, accurate and timely knowledge exchange and choice making. Accordingly, the main challenge in discovering a 'fit-for-purpose' ontology is overcoming the extant diversity of erroneous or conflicting mental models. A business sans common ontology is somewhat like a middle school orchestra playing Brahms --- Brahms loses.
2) Application software is an executable model of a business activity. Particularly in object technology a class library is quite similar to an ontology of the respective business domain. It was this premise that motivated the semantic mapping approach. Good thing we didn't have lots of Smalltalkers.
OBTW, I do not claim that our employer was pleased (they were billing man hours not class libraries). However, our customer was.
Jack Ring
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2011/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2011
Community Portal: http://ontolog.cim3.net/wiki/
-- Michael Uschold, PhD Senior Ontology Consultant, Semantic Arts
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2011/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2011
Community Portal: http://ontolog.cim3.net/wiki/
|