Arithmetic units

CB,

Sometimes called the FOCUS number system, there is plenty written about this and even some logarithmic DSP chips for sale.

All arithmetic (except integer) is inexact, e.g. adding two numbers together where one is much smaller than the other, something must be lost with a limited amount of space to represent the result.

In most floating-point, addition is exact provided that you can represent the less significant bits, but multiplication produces much that must be discarded. With logarithms, exactly the opposite is the case, as multiplication and division produce exact results, but addition and subtraction are inexact. The inexactness is in the "fudge factor" table whose entries have been computed from transcendental functions that must be rounded off to the precision provided for.

My original point is that this would all be obviously impossible if you believe the college profs who say that you can't add with logarithms. In short, what good is a math education when the stuff they teach is just plain wrong?

Years ago I took the CDP (Certificate in Data Processing) exam. If you were to research and answer each question perfectly correctly, you would probably not pbutt because of all of the "conventional wisdom" in the exam. Most of my "study" was learning the current (and dysfunctional) conventional wisdom. Now, this is creeping into other exams, so that children aren't so much demonstrating their education, but rather they are demonstrating their degree of "programming".

Steve Richfie1d