Thursday, July 12, 2007

imho Ocean Semantic.............._


Ah, coast sweet coast.

Hey, the coast is clear for me. I think I'm clear on what Semantic Web, semantic interoperability, ontologies, taxonomies, folksonomies, and all that are all about, and why they're important. But not everybody is clear on the concept of concept-based data modeling, search, integration, etc.

And the people who still don't see the point of it all are in the vast majority of the IT world. They're not stupid, of course, but are justifiably skeptical of a semantic wave that's been taking far too long to cross the SOA ocean.

Case in point: Bill Inmon, the guy who continues to be the father figure and best-practices prophet for the data warehousing universe. Check out what Bill said last month: "I admit it. When it comes to semantics, I don’t just get it. You can call me misguided, an old fuddy duddy, or just plain dumb. In one way or another, perhaps all of those names fit. But at the end of the day, I just don’t understand semantics."

Let's be clear on what Bill was saying. It's not that he doesn't understand what semantics is or is not--it's just that he fails to see what value-added the "Semantic Web" adds over and above traditional approaches to semantic integration. Inmon at length:

"One branch of semantics I looked into with great interest was ontologies. Having done some cursory work in the field in my own software development, I thought that surely here was a value proposition. But no. At least, I couldn’t find it.

"Then I looked at semantic logic. Now semantic logic is quite interesting. It reminds me of a really good crossword puzzle – the kind I like to take on flights between Europe and the U.S. But while semantic logic is interesting, how it applies to any business problem is beyond me. No luck here.

"Then I looked at linguistics. This was perhaps the biggest disappointment of all. Linguistics has been around for years. There have been countless hours of research and countless government grants in the field of linguistics for a long time now (at least 30 years). And where is the business problem that is being addressed by linguistics? Certainly it is nowhere on a large scale. It is true that there are some small startup efforts that make use of linguistic technology. But after thirty years of research, you would think that there would be a lot more technology on the table – a lot more proof in the pudding."

I quote Inmon at length not to agree with him so much as to point out the degree of intellectual fatigue implicit in his comments. Immersing yourself in this sometimes shiny, often briny, occasionally nonspecific ocean of academic and technologic goodies can be suffocating. I chose to test the waters simply because I realized that Semantic Web is slowly inundating the coastline in EII, ECM, BI, DQ, MDM, ESB, and other of my core coverage areas. Also, further investigation of this space has been one of my back-burner personal/professional/intellectual to-dos since 2004 when, in another professional context, I was actively discouraged from going there. Had to go there. 'Twas quite good, this time around, that my podcast colleagues (Dana Gardner, Dave Linthicum, Tony Baer, Joe McKendrick, et al.) at were quite agreeable, amenable, and ideational on this topic, and chose to go there with me. Linthicum, in particular, is an expert himself on the matter, and was a major inspiration and launching point for my random-swim through this specific sea these past two months.

From a blogfodder standpoint, one of things I like about Semantic Web is that one can pretty much connect it to anything one wishes on a philosophic, ontologic, metaphysic, technologic, or sociologic level. What a wonderful context, all this semantic, for this old boy to wax pedantic. I almost wanted to truck into it all a meditation on Immanuel Kant's Critique of Pure Reason, the distinction between "things in themselves" (i.e., true ontology--the study of the objective ground of reality and being) and "things as they appear" (i.e., phenomenology--the study of the subjective apprehension of reality and being)--and to note that Semantic Web "ontologies" are in fact simply an "a posteriori" analytic practical prescriptive institutional instrumental consensus on "things (e.g., subjects, predicates, objects, entities, instances, classes, groups, relationships, attributes) as they appear" to a semantic governance authority within some semantic domain--not actually the sort of "a priori" synthetic metaphysic revelation apprehension of the ineffable root of things that Kant said he had no problem accepting as a possible article of faith but saw as essentially a futile field of unverifiable undecidable propositions outside the sphere of pure reason, which Kant essentially aligned with scientific process of progressive societal construction of an interlinking system of empirically verifiable statements through the building and testing of interpretive frameworks through controlled observations--almost wanted to note that the linguistic concept of "semantics" is a subset, not the sum total of "meaning," since language is an instrument of human intention and any statement can be understood as much in terms of its "pragmatics" (e.g., the explicit or implicit, interior or ulterior motives and objectives that shape and direct it) as it is in terms of its "semantics," which concern the role of language as depicting objects in the objective/real and/or subjective/imaginary world as worthy of our contemplation, and of course, this same language can also be understood as conveying meaning on the level of "poetics," which doesn't so much mean you have to get into William Carlos Williams or Wallace Stevens but rather have to realize that language is itself an object worthy of our contemplation, and that language, this object-in-itself, is memorable only to the degree that the concocter of said bonmots has consciously or otherwise wielded all the communicative devices of good composition, parallelism, grammar, wordchoice, alliteration, tintinnabulation, emotion, tone, guided imagery, etc to make the language itself, and hopefully the semantic objects and pragmatic objectives contained therein compelling and worth our while dwelling therein and upon--but that would be pedantic.

Hey, check out the work of the W3C Emotion Incubator Group ("a general-purpose emotion annotation and representation language). And consider the growing text-analytics-industry implementation of semantic "voice" (e.g., "Voice of the Customer"--see what Attensity, Clarabridge, and others are doing in this sphere) , which is essentially something capable of being extracted as a component of the human "sentiments," moods intentions and propensities, that lie latent in unstructured colloquial text.

In the Social Semantic Web, meaning may be measured in the meandering moods of masses meditating on mashups.

All for now.