Friday, January 13, 2006

The role of math

BusinessWeek has a cover story article in the January 23, 2006 edition by Stephen Baker entitled "Math Will Rock Your World" which reports on an alleged "math revolution" revolving around the use of powerful mathematical techniques for analyzing business data. As far as I can see, there isn't much new here. The use of math for analyzing volumes of data has been quite common for quite some time.

I suspect that the real news should be not how much can be done with math, but how little of durable value has been accomplished with so much effort and expense.

And the big, yawning issue will remain how well managers and executives really comprehend the math upon which the operations of their organizations and their very decision processes depend on techniques which may be beyond the comprehension of all but a very, very few elite technicians.

One of the most critical issues will always be the set of assumptions which are made whenever real-world non-numeric data (e.g., text, customer opinions, etc.) are converted to numeric form. That conversion is almost never a clean, error-free process. How many managers and executives really have a handle on what those assumptions might be? And the assumptions get very, very complex once you get into fuzzy logic, one of the most important approaches that will yield important results in the future, but isn't even mentioned in the BW story.

The article doesn't even mention the Semantic Web, ontologies, or a lot of other important research that will drive the organization of data, information, and knowledge in the coming years.

The core problem here is journalistic bias. A couple of times the BW writer has let slip that he believes there are two kinds of people: word people and number people. He has failed to recognize that there is at least a third important category: symbol people. They are the ones who will be figuring out how to make the Semantic Web, ontologies, fuzzy logic, etc. really work well in the coming decades.

The bottom line: math was the story five, ten, fifteen, and even twenty years ago. There used to be a name for it: "rocket science". The BW article even referred to a term that was quite popular a number of years ago: Quant.

Yes, math is still around and is still worth covering, but as a cover story that seems to imply a breakthrough? That makes no sense.

Symbol manipulation, including concepts, ontologies, the Semantic Web, etc., will be the big story going forward, except for journalists who don't have a clue.

To be sure, there is also the intersection of math and symbols and quite a few of us have worked in that mixed world for many, many years, with little recognition. Yes, math does have a role, but we should be careful not to overplay that card.

-- Jack Krupansky


Post a Comment

Subscribe to Post Comments [Atom]

<< Home