From: Chuck Brotman <chuck.brotman@...>
Date: 2005-07-14T03:19:06+09:00
Subject: Re: Updating GUIs


"Martin DeMello" <martindemello@yahoo.com> wrote in message 
news:jQ4Be.1947450$Xk.771377@pd7tw3no...
> Joe Van Dyk <joevandyk@gmail.com> wrote:
>>
>> Oh!  That's a great idea.  I understand now.  The cost of that is that
>> I need 72 times more memory for the TkcPolygons, right?  Which
>> probably isn't too significant.
>
> If it is, a halfway measure is to precompute the sines and cosines of
> angles in five degree increments, and get them from the lookup table
> instead.
>
> martin
>
>

I agree with john's post.  I'm no expert on the object oriented aspects of 
performance (object creation, etc), but I'll try to help from a purely 
"functional standpoint. It's Hard to tell what your bottleneck is, without 
seeing a performance profile. but, from my experience (with c++) the 
transcendental functions (sin, cos, etc)take a significantly amount of 
computation!  Caching pre-calculated results for polygon rotation or even 
just the sin/cos functions can make an order of magnitude difference, and 
don't really take up that much space. (I did it for 1-degree increments and 
had no problems.  Also, even in a faster language, like C++, you may need to 
be careful of  unintended integer/float conversions, especially in the 
middle of functions.  If you can get everything in terms of integers you can 
get much improvement.