Friday, April 07, 2006
The Second School of Semantic Science
ontologyMapping Glass Bead Games
Focus on the Reference Model for Open Distributed Processing (RM-ODP)
and
related discussion about views, viewpoints and community centric
Continuing
discussion on the SOS CoP forum
It was suggested:
I
think in either the approach you mention or what I'm trying to suggest, the
most important thing is:
start from the business requirements
After
that, you can attack the implementation in whatever way makes sense.
I agree in principle, but one has to be sure that the business requirements reflect true market needs.
The first step should not be about the business requirements. It should be about shifting the origin of design over information structure to non-IT centric professionals and everyday humans. The IT business will be rewarded still, it is just a question of who is in control.
This means bringing knowledge management practices from collaborative spaces like weblogs and virtual communities (which are older than the URL (1991)).
The Business Centric Methodology (BCM) is discussed at:
www.businesscentricmethodology.com
I found it late and feel fortunate that David and others advanced this starting in 2003.
My work brings together a number of other people's work (as discussed in the BCNGroup Roadmap) to provide a
1) measurement of the conceptualization of individuals (as in a conceptual (semantic extraction) process from the written text in weblogs) as discussed at
http://www.ontologystream.com/area1/MemeticOntology/mappingSocialSymbols.htm
the conceptualization could be by UML, OWL, Topic Maps or any other means including simple expository text or story telling.
2) this individual conceptualization corresponds to the bottom of four layers of the BCM specification
3) the individual conceptualizations are then aggregated (and this was specified in the Roadmap and will be refined and specified in the Community Centric Service Methodology specification) into a "community" viewpoint.
The key is the notion of a choice point, and this is where the design of information structure (so called "semantics") needs to occur... not at first by the information engineers. (This is the difference in approach to SOA that we are integrating). Let us demo that. (smiles)
I note the excellent work on RM-ODP
http://www.lcc.uma.es/~av/wodpec2005/papers/3-multiview.pdf
where view and viewpoint are separated so that viewpoint has the power to be different from other viewpoints.
I also note the excellent work at Rosenett and at CoreTalk.
The work on the "Pragmatic Web" is excellent:
http://www.starlab.vub.ac.be/staff/ademoor/papers/iccs05_demoor.pdf
this point of view is "second school", in my opinion. (Smiles)
the other two BCM layers are for "extension" and "implementation". This is where OMG and other business groups should be working - but unless the first two layers are done (outside of their control) then the complete BCM is not done right (again, in my opinion). If the other two layers are done then more than sufficient work will exist for OMG members (if they are willing to refactor what they have been doing so that ordering is correct).
The issue is where the origin of design occurs.
The notion that the first thing is to focus on business requirements misses the point entirely. One cannot know the useful things that business might do to serve the public good if the business is too business oriented? The monkey whose hand is in the cookie jar is a good metaphor.
The suggestion is that this business centricness and in particular the IT-business-centricness over exploits the market and as a consequence we all suffer with dysfunctional IT systems (ie no interoperability, no agility, no reuse, no service).
The alternative that is being set up with the new (Community Centric Service Methodology) initiatives is that SOA be seen as supporting a mapping of the transaction spaces FIRST, and then let the businesses serve the public. (smiles.)
Many are willing to do this, what we have not had is leadership.