Civileme wrote:
> <snip>
> It is relatively simple to prove that the most efficient method
> of storing information for lookup is to have it sorted (or where
> the individual information objects are large, to have a sorted
> index) and to search the index on a binary basis--that is, look
> in the middle and go whichever direction the mismatch indicates,
> up or down and look in the middle of what's left. Using a binary
> search algorithm, while you think of a number between 1 and 1000,
> I will start with 500 and say 250 n3ext if you report "high" or
> 750 next if you say "low", and I WILL nail your number in 10 or
> fewer guesses if you just report whether I am high. low or
> matched. The problem with a binary search is that you HAVE to be
> sorted. So updating your database with a time-eating sort had to
> be done when the tail log of additions grew to the point that the
> lookup became inefficient. A balanced tree duplicated the quick
> lookup capabilities of the binary lookup (nearly) by making nodes
> of sorted information with each node about half-full of data, two
> new nodes being created from a current one when the node filled
> to a certain point. Now this wasn't done in a single level but
> in a tree structure. This entity is called a Balanced tree,
> B-Tree, B+Tree, etc, and a short book could be written on the
> code necessary to implement it. It is about six times as
> complicated as competing storage algorithms.
>
> <snip>
>
> Civileme
I used to program databases in that OTHER OS and kept running into a
package called B-trieve and always wondered how they came up with that
name/method. Thanks.
--
Joseph S. Gardner
Senior Designer / Technical Support
Kirby Co., Cleveland, OH
[EMAIL PROTECTED]
The box said "Requires Windows 3.1 or better."
so I got Linux.
Registered linux user #1696600
ICQ #63389227
