Description:
<jats:p> Information system analysts often say that information reduces or even eliminates uncertainty. This is supposed to be a good thing because uncertainty disagrees with the nervous system of executives, who recognize only two possible probability values: 0 and 1. That information reduces uncertainty is “scientifically proven” by invoking the Wiener-Shannon theory of information and the concept of entropy. </jats:p><jats:p> Now I find all this confusing or downright erroneous. Perfect information or complete knowledge—difficult, expensive, or well-nigh impossible to find—does eliminate uncertainty. But imperfect or little knowledge—a more practical commodity—does much more curious or subtle things. </jats:p><jats:p> One of the aims of information calculus advocated by this writer is to clarify and remove confusion in this field. So in this column I illustrate what imperfect information does, and show that the phrase “remove uncertainty” does not make good sense. </jats:p>