[futurebasic] Re: [FB] New, speedier INDEX$ arrays

Message: < previous - next > : Reply : Subscribe : Cleanse
Home   : August 2004 : Group Archive : Group : All Groups

From: Scott Spencer <sj3@...>
Date: Mon, 30 Aug 2004 09:56:35 -0700
tedd,

On a related note:  Have you ever seen or check out "Linked Lists 
v1.0.0" by Richard P Phillips?

It looks to be a very good FB Library with some excellent documentation.

On Aug 30, 2004, at 7:14 AM, Jay Reeve wrote:

> On Monday, August 30, 2004, at 07:29  AM, tedd wrote:
>
>> Jay:
>>
>> Yes, please send me your new index.
>>
>> One of the things I've been thinking about is a set of functions to 
>> do a similar thing.
>>
>> As you know, I'm big into linked lists, which are Okay for certain 
>> things. But, a few months ago I was studying a C++ book wherein they 
>> used a linked list -- but hidden from view they inserted a red/black 
>> (parent/child) director which organized the list into a b-tree. All 
>> of the b-tree stuff was hidden from the user.
>>
>> Considering such, I thought that a set of functions similar to INDEX$ 
>> could be written that would handle operations without the user 
>> knowing that they were using a b-tree.
>>
>> As you know, the look-up time is hidden in the initial "how the data 
>> is stored" -- if it's stored better, then look up time is less.
>>
>> It's been my experience that a b-tree approach is the fastest 
>> algorithm I've seen for looking up data. In the worse case scenario, 
>> it's "time to find stuff" the same as a sequential linked list lookup 
>> (only one limb, not balanced), but with a balanced tree, nothing can 
>> beat it. What do you think?
>>
>> tedd
----
Scott Spencer
sj3@...


--
To unsubscribe, send ANY message to: futurebasic-unsubscribe@...