Saturday, May 21, 2005
The next generation knowledge tools
Extended comment on the 80-20-deployment effort ratio.
The Knowledge Foundation Inc concept has a long history. Richard Ballard has had success in getting 7 or 8 government contracts and completing them with client satisfaction. Moreover, his theoretical foundation is well defined and has sufficient grounding in science. So it is clear that Knowledge Foundations might be the next step in how sets of concept representations are used by society and by government.
The current Semantic Web school of thought, what we have called the “First School of Semantic Science”, has some characteristics. Everyone knows what these are:
1.0: it is deemed that ontological reasoning is important
1.1: reasoning involves things like a formal process called consistency checking, where for the system to be consistent two incompatible answers cannot be obtained.
1.2: reasoning involves checking to see if the process of consistency checking can always be carried out in a reasonable amount of time.
1.3: reasoning also comes to mean that the computer program does not crash due to some software bug.
2.0: it is deemed that the universal resource indicators (URI) is to be used to standardize meaning of some word or string or words.
I stop here, because I am hoping that I have fully represented the First School’s positions.
In both cases, the Second School finds technical and science-based faults with the First School. We have addressed this elsewhere.
I wish to return to the discussion about Knowledge Foundations Inc and the KFI use paradigm that we see.
We have said that the 80-20 ratio exists where 80% of a second school deployment effort has to be made by the “client” and only 20% can be made by “contractors”.
Objectively, this is what everyone except the contractors would want. But such a 80-20-deployment effort is not what is expected and thus has to be both explained and illustrated through successful deployments.
The Roadmap was designed to bring a 1.8 Billon dollar investment by the Congress in modernizing US Customs and Border Patrol’s information technology. Specifically we sought to use OntologyStream’s 90-day contract to identify a working paradigm for CBP that would provide Informational Transparancy in Secure Channels over all CBP inspections of cargo containers.
The subcontract cost of the integration and deployment has been set at $750,000 for an initial operational system integration and field deployment.
The proposal is basically ignored by the Lockheed Martin subcontracting officer, and has not been fully communicated within Custom’s directly.
An alternative proposal might be based solely on KFI’s Mark 3 system.
However, the key elements to the Roadmap are semantic extraction processes (Readware, NdCore, and SLIP tool sets) that produce results that are compared with explicitly defined ontology written as RDF files (using Intellidimension Inc’s RDF repository tool set). Two knowledge management system operate independently but in a collaborative fashion to build and maintain managed vocabularies and a mapping between word based subject matter indicators. So there are three components:
1) ontology persistent repository and standard ontological models (Intellidimension Inc using RDF)
2) semantic extraction processes (Readware, NdCore, SLIP)
3) knowledge management connecting user community use of language with ontological models
Given that the problem we are solving is of some important; is there a natural way to integrate KFI in with the Roadmap, or should the Roadmap be dropped and we use only the Mark 3?