Community centric service methodology à [home]

Conversation between Dr Roy Fielding and Dr Paul Prueitt

Previous discussion à [48]

Roy,

I am a friend with John Sowa and Richard Ballard. In broad terms, what they advocate and work on is an n-ary representation of a concept or something that could be a concept. John feels that common logic should be standardized to place an inferential framework around the output of n-ary production.

John's opposition to RDF is complicated to understand, but I feel that you and John would agree if the real constraints, on the RDF triple, were communicated between you and him.

Richard has given up on inferential logic completely and uses an GUI to "hand code" topics and links between topics. His work has been funded by DoD for thirty years. It represents a revolution in the making (www.knowledgefoundations.com) but my sense is that there is too much hype related to the notion that machine can reason like humans do. As John has pointed out an number of time, to Richard and others; the Knowledgefoundations Inc may become the next Cyc Corp; but may fail in stated objectives because of the sales pitch. The underlying work is however an advance over RDF, in my opinion.

This nary representation comes directly out of latent semantic indexing and related techniques - which is my highest level of expertise.

The means to link raw "co-occurrence" patterns to RDF and then to OWL has not been explored (that I know of), except in my Roadmap and other (obscure) efforts. My technique is about 80% completed in work I have done over the last eight years - working on pattern identification and the use of patterns to "extract entities".

I did not really choose to have a hard problem, but this is what I have and what drives me. (smiles). I am looking for some way to make a transition to just working, not trying to find funding.

I have some proto-types (advanced in function but no use model).

This gives me pause to believe that my work is close to being completed... in a way that "provably optimally" addresses the real time interactivity between humans, human communities and software agents. So this is the big one. The big one allows the shift of control over distributed computer processes form the IT sector to everyone … ubiquitous computing that is not owned. Many people believe that we get ubiquitous computing and someone, themselves, are the owners. This is both morally and socially wrong; but also incorrect form a technical point of view (a view that I am well known in stating as in www.secondschool.com)

Now I know the work I am envisioning, with some correcting and modifications, would solve some engrained problems... I just have to get to the point where my work can be supported and directed for six months to a year.

The above note has a lot that can only be understood by someone like your self who has been in the inner workings of the current Internet system.

I have one on going project now in "services".

I have two or three projects planned (one in global Cyber Security with faculty at Norwich University) and two private; where a provable secure ultra stable operation system for "agents" / "services" / "transactions" would take my work (also as extended by two "students") on n-ary processing ..

http://www.bcngroup.org/area2/KSF/Notation/notation.htm

and bring this to the next step. I would prefer to use Forth, but realize that Java with the Spring framework is more likely .. The "key-less hash table" data structure is what you would find interesting.

Sorry for the long discussion...

Your work on representational state transfer could be tied together with n-ary processing using convolutions - as discussed in the notational paper. I have prototyped this several times and have operational visual basic demonstrations - so this is theory plus demonstration.

I would like to work on a description of how REST and n-ary transforms would be "specified".

Your thesis is a good reference point.

Dr Paul S Prueitt

703-981 *&spamfilter*()()*- 2676

Previous discussion à [48]