In article <38a97257@news.povray.org>, "Nathan Kopp" <Nat### [at] Koppcom> 
wrote:

> This would be more memory efficient, to be sure.  It might be slower 
> (due to more render-time transformation calculations), but if done 
> properly, it shouldn't be too bad, and the option could exist to 
> expand all references for speed reasons if the computer has lots of 
> memory.

Hmm, although the transformations would slow things down slightly, they 
might prevent some people from having to go to virtual memory, so it 
could in some cases be faster.

Now that I have thought about it more, it shouldn't be too hard to 
implement even in the C source. Just add a flag to each object 
specifying the storage method, maybe add a counter of objects using the 
data, and make some small modifications in the functions which copy and 
destroy the objects. Make sure the data isn't destroyed if an object is 
using it. And of course modify the object data structures and the code 
which uses them so they can share data.
I don't really have the time to look at the source right now, though, so 
I am just going by memory...I am too busy with my particle simulator, a 
couple other potential projects, learning the MacZoop application 
framework, thinking of learning more Java, and of course school. :-(
I am beginning to wish days could have 36 hours...although I am sure 
school would just take more of them...

-- 
Chris Huff
e-mail: chr### [at] yahoocom
Web page: http://chrishuff.dhs.org/

Post a reply to this message