Archive for September, 2007

Alex Iskold recently authored a couple of articles over at the Read/Write web about the Semantic Web. I caught the one labeled Top-Down: A New Approach to the Semantic Web that refers to an earlier post on the same subject. I commented at the site that I thought it was misinformed and not very helpful at all. When I say it is not helpful, I mean that I do not believe it is helpful to the community, to progress or understanding.

I must say upfront that I never met Alex Iskold, but from his own bio and articles I have read, he appears intelligent and accomplished. I am not picking on him and I am not attacking him personally. Any interpretation of what follows to the contrary is just wrong.

I will admit also that I am holding the Read/Write Web to an editorial and perhaps even an academic standard while I recognize the hosts are not obligated to any such standard. Nonetheless, I am compelled to point out that neither the author or the hosts are authorities on the Semantic Web.

This leads us directly to what I have issue with, which is: some authoritative posturing that is both misleading and detractive. So this post is my rant against all those who hype semantics without much knowledge about semantics proper, no experience with ontologies or graphs and certainly no semantic methods or algorithms of their own. I am only using Alex’s article as an example but the SEO scene is overflowing with this kind of noise that proliferates confusion and ambiguity and leads to misunderstanding.

First, let us define what I called semantics proper. For our purposes we can use this definition from Wikipedia: Semantics (Greek sēmantikos, giving signs, significant, symptomatic, from sēma (σῆμα), sign) refers to aspects of meaning, as expressed in language or other systems of signs.

A sign may be expressed in a single word. A signifier of explicit command and control is often expressed that way: stop, go, turn. The word ‘stop’ is a linguistic sign. It is used to indicate or point out a specified meaning: that meaning is to halt further motion. There is a clear and unambiguous relationship between the giving of the sign and the effect to be “understood”. Its purpose is also clear: to command and control. It is clear, in this case, that the aspect of meaning being signified or being given as a sign is not the same thing as the linguistic sign used to express it. Please try to keep this in mind.

Semantics are measures used by people to interpret significance in the real world. In this sense, signs are the guideposts of thinking about, communicating (collaborating or corresponding) and relating things. Another way of saying it is that meaningful or significant aspects of interaction are expressed as words, as nouns, verbs, and as subjects (AKA linguistic signs) in any language, but they are not the same thing as the linguistic signs used to express them.

In Alex Iskold’s post, he is using words in ways that appear significant, but upon deeper examination, or under a critical eye, are not. They are as I have charged — just noise. Why? Because of the way he chose to characterize the Semantic Web and the actual fact that he offers no new approach at all.

In language and in human discourse, names, used as indicators of something, must be characteristic at least and obviously symptomatic at best. When they are not, the semantics (semantic measures) must be called into question. I will come back to this in a future article; for this post I just want to substantiate my charge that Alex’s Iskold’s two posts at the Read/Write web are just noise. Critical thinking is all we need in order to achieve a clarity of understanding of the significance of discourse on cultural and social institutions (such as the Semantic Web).

The Semantic Web of Sir Tim-Berners Lee (TBL) is not about enabling “computers to ‘understand’ semantics the way humans do” as Alex Iskold claimed in his article about the difficulties with the classic approach. This statement is depreciative because it is completely counter to what the Semantic Web is really about.

I would first suggest that most humans, people, do not understand all this discourse about semantics, and are rather quite confused when the discussion ranges into semantics. Yet, people use semantics proficiently in all of their interpersonal relationships. And groups of people, indeed, entire nations rely on their understanding of semantics. This is necessary in order to realize that the threat of the use of WMD by Iraq was a ruse rather than a reality, for example.

The objective should be for we people to understand and articulate these semantics well enough that our machines may process (use) these semantics to better relate and collaborate with us, and to help us identify and discover relations in discourse and between languages, cultures and society at large. But that is also another story that I will cover in another post.

When people agree that some thing or topic of discussion is a matter of semantics, they tend to drop the subject or topic of discussion. The reason being: there is a disagreement over the semantics (of naming). It is understood that nothing comes of that (disagreement over the naming). So in fact, humans do understand semantics very well

So since people are not confused about using their sense of semantics in interpersonal situations, I am beginning to ponder whether this purported lack of understanding is some sort of occupational hazard. Software architects and computer engineers, it seems, tend to be unsocial if not altogether inhuman in the face of the computational complexities of sense-making. So the idea has some merit on the face of it.

Because some of them can assume authoritative roles, it then creates problems in understanding because the noise they proffer muddies the waters and creates barriers to understanding. When popular publications are driven by monetizing every story they post, they seem to be more willing to compromise their editorial integrity and allow this noise into the blogosphere and cyberspace. This tends to a more serious problem where “information” becomes so diffused and ambiguous that it is no longer possible to meaningfully communicate except in small close-knit groups.

Fortunately, a little exegetical examination can help people spot those problems. Let me demonstrate how one can apply critical explanation to the first paragraph of Alex’s article:

“Earlier this week we wrote about the classic approach to the semantic web and the difficulties with that approach. While the original vision of the layer on top of the current web, which annotates information in a way that is “understandable” by computers, is compelling; there are technical, scientific and business issues that have been difficult to address.”

The first sentence requires one to read the previous article to substantiate the so-called difficulties of the approach. I will come back to this first sentence after looking at the second. Let’s just take the phrase ‘which annotates information in a way that is “understandable” by computers‘. Not considering whether that was really the “original vision” of Sir TBL; think about it: Ask yourself: Are there computer programs that annotate their information in ways that are not understood by the programs in those computers? Does that make sense?

This is what semioticians call a sign, a significant hint, a semantic indicator that, in this case, signifies something is amiss. I do not really care what linguists call it or whether they agree, because we are not talking about parts of speech or sentence construction here. We are talking about the significance of the words found in a specific text and discourse and how those words capture, divide and signify a specific state of being. Is the article really about the Semantic Web, an alternative proposal, or is it just a ruse?

While Alex might rightly complain: that is not what he meant at all! This example explains why we need to use exegesis so that we might begin to benefit from “understanding” what is really happening. Because it is tedious and requires close examination it would be a good task for computers. Because people are not always prone to a critical examination and tend to be far more superficial, it answers the question of why we need computers to understand semantics. Okay, now let me correct Alex’s first incorrect premise by offering some evidential resources for you.

Alex Iskold’s claims about the Semantic Web of the W3C are empty because they have no relation to Sir TBL’s vision and are not in the least representative. No matter how sophisticated or insightful Alex’s top down approach is, the premise he gives does not support it in the least. It is really nothing new as well, but rather describes the status quo and the way designers architect their systems today. Alex even gave examples of a few companies structuring data in one way or another.

Sir Tim-Berners Lee’s vision was for a universal means for computers to store, link and exchange data. This vision has nothing to do with human knowledge except as a by-product of organizing data that may (in turn) represent human knowledge in more uniform (and interchangeable) ways. It is certainly not about enabling “computers to ‘understand’ semantics the way humans do”. Anyone with any doubt can view this rather recent video of Sir TBL’s testimony before the U.S. House of Representatives, Committee on Energy and Commerce, Subcommittee on Telecommunications and the Internet.

In order that you do not have to watch the entire 2 hour plus video, here is a time index of relevant parts:

  1. Sir TBL begins his testimony about 30:00 minutes into the video and discusses his vision —the universal nature of the world wide web– until about 37:00 minutes into the session.
  2. He begins discussing the importance of data integration at about 41:50 minutes.
  3. He wraps up with why it is all necessary from about 44 minutes through 49:00 and continues his testimony until the time index of 52.05 minutes.

After that Sir TBL takes some questions, while the U.S. representatives jockey for some sort of platform of theirs, addressing each issue as the great gentleman that he is. I would encourage anyone to listen to the entire session.

Not that I relish supporting the Semantic Web, without a semantic theory, those who take the time to listen to Sir TBL will agree that Alex Iskold cannot possibly be an authority if he is this confused about the vision for the Semantic Web in the first place. I mean; it is clear for example, that the foundation (the internet) and the infrastructure (the weaving together of standards, protocols and programs that underly the web) make way for the ground-breaking applications that will free data from inaccessible information silos.

Alex suggests we maintain the status quo, store data in ways only a few can use, lock us into a structure and an approach– just as RDF locks us into a single and very specific semantic form. Let’s see… Alex’s top down tree — or — the standards of the W3C. Which authority would you choose?

The Romans used to say that all roads lead to Rome. We citizens of modern civilization can only hope to free our authorities from the trappings of top-down and bottom up ideas and approaches to the publicly funded citadels and depositories of the knowledge of human culture. What to do about those who try to pass noise off as authoritative opinion and commercialize the affair is another matter entirely.

Please let me know what you think.

Read Full Post »