Pixel Invariance

[Simon F]

Simon, I'm shocked. You'll be saying you haven't seen this one next...
http://cop.utmem.edu/lcleavelin/PLfoot.htm

Oh I've definitely seen that. There was another that compared languages to different transport systems.

 

[davefb]

Isnt the issue about 'foating point compares' similar to experimental accuracy . IE , you shouldnt code "if a==0.1" but code "if ((abs(a)-0.1) < ERROR_LEVEL)" instead ?

Grief my physics and maths teachers really must have rammed the point home about values being useless without the errors hehehe .

Obviously i really dont know how that applies to the comments about renderers . . it was more for the general coding .

-dave-

 

[Dio]

Actually that was the point of the whole thread. I think the argument is that bit-precision is unnecessary for 3D renderers...

 

[akira888]

you shouldnt code "if a==0.1" but code "if ((abs(a)-0.1) < ERROR_LEVEL)"

Excuse me if I'm wrong, but shouldn't it be -

if ((abs(a-0.1) < ERROR_LEVEL)

- instead?

 

[davefb]

i'll get my coat,,,

fwiw, i'm working on symbian which compiles for the 'simulator' using vc ( into x86), then for the hardware compiles using gnu to arm. one has hardware fp, the other doesn't . so this kinda thing does have the possibility of biting me in the ass.

-dave-