Wednesday, January 18, 2006
Additional reading:
Cory Casanave's paper on Data
Access
work on ontology for biological signal pathways
[131]
ß parallel discussion in
generative methodology bead thread
question about the extend of use of the eBusiness standards
Earlier
part of this discussion
[342], [343], [344], [345], [346], [347], [348], [349], [350]
Allen
Your note posted at:
http://colab.cim3.net/forum//ontac-forum/2006-01/msg00052.html
goes to the heart of how web services might be mediated (soon) by the new OASIS WSML (Web Service Modeling Language) and WSMO (Web Service Modeling Ontology) standards.
The intentionality of social interaction is often not determined with the type of precision that is demanded by W3C and Semantic Web theory.
However, if one gives up the current SW approach there are specific alternatives that have been well developed but not widely seen. The Adi ontology is one of these. There are others (Knowledge Foundation's Mark 3 nary processor and many many semantic extraction internal ontologies. In most cases, these are developed from a linguistic foundation (at least) and not a hard formalism. Some times the approach is "semiotic", in nature. In most cases they (these "semi-formal" ontologies) are sets of concepts, unencumbered by classes, class hierarchies, and the now hundreds of varieties of "logics".
An interesting paper is the recent Dagstuhl Seminar "Semantic Interoperability and Integration: Executive Summary"
http://drops.dagstuhl.de/portals/index.php?semnr=04391
I feel that there are two themes in these papers
1) an discussion of the intractability of the effort at establishing a
"mathematical foundation for semantic science"
2) the interest in preserving this effort by developing community based agreements that are directed (not by a strategy to give society a new set of capabilities based on high fidelity flexible ontological modeling, but) at disguising the failure of the current approach and institutionalizing knowledge engineering as an advanced esoteric discipline that no non computer scientist could hope to comprehend. They follow the lead of the AI community, a community that has well established itself in the university system and which maintains its funding and status in spite of almost universal acknowledgement of its failures and absence of a credible foundation. [1]
The alternative is to build a foundation for ontological modeling that is consistent with natural science and everyday social and mental activity. The "knowledge engineer" may feel that this is an absolutely disguising notion, but there it is.
Imagine, ontology that can be understood by the
user!!!!!
By joining the Protégé eforum, one can observe first hand that each day there are often one or two new individuals starting the frustrating and most often non-productive path at developing, using or just being puzzled by the Protege software and the existing W3C "ontologies". This becomes computer science for computer science sake.
Fact:
The purpose of developing a specific ontology is constrained strongly by the limitations of the tools at hand. In the case of the ONTAC working group, the group can move through a full several month's long discussion about the limitations of first order logics, and then merely ignore the eventual partial conclusions.
I was thinking today about the fundamental modeling issue. This issue should be addressed by defining a categorical difference between natural systems and formal systems, as I did at:
http://www.bcngroup.org/area3/pprueitt/kmbook/Chapter2.htm
In resent discussions, Judith Rosen has developed some summaries of the proposed definitional distinction between "simple" and "complex". [127]
A group of individuals are making the point that any formal system taken literally is non-complex, even if the size of the “system” has become very complicated. If nowhere in the simple formalism is there an appeal to human induction, or induction via real natural mechanisms embedded fully in time, then the large more complicated systems never produce an emergence of a whole that is in any way greater than the sum of the parts. “Emergence” does not occur, in that case.
The OASIS WSML and WSMO standards are attempting to leave open the imprecision of matches between the request for complex web services and the response (where complex web services are actually assembled and delivered). Our problem is that the language (terminology) we need to use,
e.g. “expressiveness”, “semantics”, “ontology”, “intention”, “complex” etc,
is considered “owned” by the SW and AI community. To not agree is to be accused of simply being ignorant of what the “truth” is.
It really is the case that “they” feel that this ownership can be and will be enforced. Over the past five decades, “they” have been right.
Fractals are a good example, and the suggestion that we make is that a fractal remains non-complex, until perceived by a human. Judith Rosen [127] talked about how both mathematics and natural language ARE complex system when seen as part of the living history of humanity.
So there is the sense of disembodied mathematics and disembodied logics.
I wonder if anyone in the ONTAC forum will make a constructive response to Ken's thoughtful and detailed description of Tom Adi "substructural" (my term) ontological framework.
http://colab.cim3.net/forum//ontac-forum/2006-01/msg00060.html
Dr Paul Prueitt
The Taos Research Institute
Taos New Mexico
[1] It is our belief that
eventually this effort will lead to legal challenges of funding to academic
centers as well as IT consulting giants under the new “False Claims” laws being
passed by the Congress.