Log Library - How is it done in the library code?

[Chengi Jimmy Kuo]

shaunc at gold.gvg.tek.com (Shaun Case) writes:

>Just as a point of minor interest, you can get the exponent of the log
>of a string representation of a positive integer by doing a strlen().
>(log base 10, of course.)

>Example:

>strlen("1") -1 ==   0   log == 0
>strlen("10") -1 ==  1   log == 1
>.
>.
>.
>strlen("1000000") -1 == 7   log == 7

>not being a math weenie, I will venture an unsure guess that the
>maximum error using this method is ~1.  If you are doing logs
>of small numbers, this is obviously unacceptable.

For this case, and you can extend this algorithm, you could use a series
of if-thens.  This way, you can determine your own accuracy and rounding.
This is a slightly different way of doing table lookup and interpolation.

strlen isn't that fast.

Jimmy Kuo
-- 
cjkuo at locus.com
"The correct answer to an either/or question is both!"