Friday, March 14, 2008

Bubble Talk about Semantics

While most of the business world seems focused on the corporate fate of Yahoo!, BBC NEWS decided that this might be a good time to report on some of their recent attempts to advance on the technology front:
Yahoo has announced its adoption of some of the key standards of the "semantic web".
The technology is widely seen as the next step for the world wide web and it involves a much richer understanding of the masses of data placed online.
The company said it would start to include some semantic web identifiers when indexing the web for Yahoo search.
The move could mean a big boost for semantic web technologies which have struggled to win a big audience.
Before addressing any details, it would be a good idea to dwell on that last sentence. Why has that "audience" for the Semantic Web been so much weaker than its champions assumed it would be? I suspect that the primary reason has to do with the never-ending opposition between simplicity and complexity. Whatever complex computations may lie behind the Google machinery for page ranking, simplicity from the user's point of view has always been the highest priority. When coupled with speed, simplicity is one of the best ways to "win a big audience:" if Google does not give you what you want on the first try, it's so fast that there is no hassle in trying a second, or even third, time. Google has become one of the undisputed masters of what many take as the primary law of user-centered design: KISS (Keep It Simple, Stupid!).

Therein lies the problem. There is nothing simple about "a much richer understanding of the masses of data placed online." Indeed, because it straddles the objective, subjective, and social worlds, there is nothing simple about understanding, itself. To reduce the argument to the bluntest of terms, how well do any of you out there really understand your mother? How many of you really understand any electronic mail or voice mail that she leaves for you? All kidding aside, if you are still struggling to understand your communications with the one person you have known longer than any other, just how much progress do you expect to make with those "masses of data placed online?" Put another way, it took Jürgen Habermas some 900-odd pages to work out a viable theory of understanding in The Theory of Communicative Action; and the reader who follows him persistently to the very end of this magnum opus may well find himself/herself "as befogged as before" (as Anna Russell put it so well)!

Once we confront these hard truths, we can see why, while the very nature of understanding can be very appealing in academic circles, there is a broad gulf between what we have learned from an abundance of academic exercises and what can actually be done in the real world. Since Yahoo! can only survive on the basis of its performance in that real world, we have to pose the classic cui bono question. Translated literally as "to whose good," we may well assume the simpler version of "Who benefits?" After all, if we can put a price on that benefit, then we can start talking about cost, which is what matters most to any self-respecting business.

Unfortunately, the BBC report is not particularly informative about the answer to this question:
At the moment most search engines, particularly Google, identify relevance for a particular topic using the interconnections between sites as much as they do the text on any single page.

The semantic web promises to change this because it helps to capture the meaning of data on a page and so give machines classifying or searching the web the capability to work out its relevance to a particular topic.

In an entry on Yahoo's blog, Amit Kumar, director of product management for the company's search site, said it was now starting to back key semantic web standards.

Mr Kumar said despite "remarkable progress" being made on how to classify meaning on webpages, the benefits of this work have not been felt by the average web user.

What was lacking, he added, was a compelling reason or "killer app" to use the semantic web technology.

"We believe that app can be web search," he wrote.
The most dangerous part of this piece of text is probably the return of the "killer app" concept. That is a phrase that was essentially blown away when the dot-com bubble burst; and there is something downright frightening about it's rising from the dead like one of George Romero's zombies. There is certainly nothing wrong with the underlying spirit of the phrase: If you cannot answer the cui bono question in terms of a well-defined deliverable with a well-defined user community, then you haven't really answered the question. The problem when the dot-com bubble was inflating was that start-ups deluded themselves into believing that whatever they happened to be doing was the "killer app;" and the only problem was to get everyone else to believe the same thing. To some extent Kumar is guilty of exactly the same specious reasoning: Yahoo! has its roots in Web search, ergo the "killer app" for Semantic Web technology will be Web search.

We are thus back in that world of reckless talk about innovation that we were hearing at Davos. This is the talk about what is new and cool (even if the Semantic Web concept has now been around so long that it is a bit of a strain to count it as either) that disregards what needs are being satisfied and what it will ultimately cost to satisfy those needs. This is not to deny that there are ways in which new software would enable our computers to facilitate our dealing with that complex problem of understanding. Rather, it is to argue against the positivist strategy of carving off a "cleanly objective" part of the problem, solving that part, and then declaring, "My work here is done!" Doing so ignores the part of the problem that resides in the social world; and, because the Internet has now become a social medium, we ignore that part at our peril, as I have tried to demonstrate in terms of the impact of "social software" in workplace settings.

The irony is that, when Google was growing to imperial size on the basis of its KISS approach to Web search, Yahoo! was promoting itself as social software: The all-purpose portal for your life in the social world. Unfortunately, this vision has not been translating into revenue with the same strength as Google's strategy of linking advertising to search results. My guess is that there are quite a few great minds agonizing over why things turned out this way. However, if those minds are too obsessed with "killer apps," they may lose any "sense of reality" of what people are doing with their computers, both at home and in the workplace; and they would forget the warnings of Ludwig Wittgenstein concerning the attempt to write a book called The World As I Found It!

No comments: