I have always wondered why they haven't. maybe it just my understanding of what going on behind the scene, but I would think that you could write some code so that it would simply take all of the dimensions and mulitply it by the scale factor that you input.
I could see things getting more complicated if it weren't a uniform scale. i.e. changing only the X, or Y and not the Z. Maybe one of you guys might have an simple explination as to why this is.