Affine transforms - matrix algebra: equals

[Pavel Safrata]

> The errbound is a double wich says that particular elements can differ 
> by its value to be considered "equal", right? Or should it rather be 
> some multiplicative coefficient?

Well, this becomes interesting when you realize that translate 1 vs. 2 
is pretty similar whereas scale 1 vs. 2 is pretty different.. But I 
don't see any obvious weighing of the elements so for simplicity I would 
stick with the maximum delta of the particular elements.