From: Mike N. 

On Wed, 20 Oct 2004 06:11:26 -0500, "Geo"  wrote:

>I don't think that's quite correct. For example how do you determine the
>least frequent character? The letter 'e' may be the least frequent in my
>search string but it may be the most common in the base I'm searching. It's
>much better imo to just search for the first character, when found see if
>the second is next, etc..

   For your application, there is no predictable character frequency.  So
if you want to get fancy, the Boyer-Moore algorithm mentioned by others
would be more appropriate by itself.

   If you are trying to pick a string out of a TCP stream, the other unique
challenge will be to span the buffers in the event the string of interest
crosses packets.

--- BBBS/NT v4.01 Flag-5