Use real number weights

I've discovered a bug with the math in the game with using vizzy output statements, and it leads me to believe this is why we have so many issues with modification of the weight.

Here's what's happening:

We use percentage to modify a base weight value. If we go less than 100%, we in facet use a less than 1 decimal number.

By using a floating point integer to first modify an original weight, we start getting smaller numbers. Once we start floating around .00004e*-7, somehow the math changes the "-7" to a "+7", and in a matter of a few milliseconds, the tiny number blows out to be in the trillions. I noticed this with the atomic timer script, where I was doing multiple calculations in one line. Breaking the line into several lines fixed the issue (also manually doing a floor by round/truncation), but I noticed through console somehow the tiny number got switched from negative to positive, why I don't know.

But that is the exact reason why I suspect that when you modify a spinning piece's weight, you go from 2m/s to 6982m/s in in 2.3 milliseconds.

To solve the issue, instead of using float point math to adjust a parts weight, just replace the weight with a full number custom input. This way you use full integers, and not micron sized float numbers that confuse the math calculations in the computer.

I could be wrong, but that makes total sense putting it all together, because it not only causes issues with spinning pieces, but I'm also seeing that it disrupts the collision detection (as in the modification of the weight of a part breaks the collision detection to where it doesn't work at all except for with the ground)