on 08/10/2000 00:01, Jay Reeve at jktr@... wrote:

>> But on my point of view this algorithm just implement hashing of strings in
>> some special order. This is used very spread in the database area for
>> indexing of strings.
> I don't really understand what is meant by hashing. I've suspected that's
> what I was doing, but would like to see a more succinct definition.

Hashing means that you at first split in some way 1,000,000 of strings you
have into N groups, then work with each group.

Usually hash-algorithm is very small and it quickly define to which group
belongs a value. 

Example: 
-- binary search for 1 million of sorted values will take
        20 steps.       ( 2 ^ 20 == 1,000,000 )
    
-- if you have hash which split into 65,535 group then after hashing you get
in average 1,000,000 / 65,535 = 15 items in group, so it needs only 4 steps
to find item.


---------------
IN your algorithm you use Hash which split all strings
into 28^4 = 614656 groups.


-------------
The simplest hash for string:

    is take ASCII value of first its letter.
    in this case we get 256 hash groups and reduce binary search on 8 steps.


-- 
Best regards,
Ruslan Zasukhin

-------------------------
Paradigma.

e-mail: ruslan@...
web   : http://www.paradigmasoft.com

To subscribe to the Valentina mail list
send a letter to valentina-on@...