Designing general rules is hard enough without confusing notions; you should start from physical principles. Why does a bullet stop? Because it has lost kinetic energy. How did it lose energy? Apart from usually minor air attrition, by piercing a hole through armour and victims; that is, by damaging objects. Making a hole requires energy, some to cause a deformation (mostly turning into heat) and some to break chemical bonds.
An object's penetrative capability is more directly related to momentum, and a formula based on momentum is both more accurate and simpler. My system is based on momentum for this reason.
While it's true that compared to other types of injury a bullet wound is very nasty relative to the invested energy, it's because, like piercing blows in general, it applies a great pressure to a small area instead of a harmless pressure to a large area, not because energy is unimportant.
For the most part, it's the open wound. An open wound causes a lot more bleeding than a closed one, so weapons that leave open wounds (anything puncturing or incisive) are much more likely to be fatal.
- How much does the bullet slow down when it passes through stuff (including air)
I was planning on a simple range increment system, where every x metres a bullet loses 0.01 penetration and a complimentary amount of energy. (Remember, energy is exponential.)
- How much the shape of the bullet degrades, affecting its penetration ability
That's all handled in DR. DD just covers loss of momentum.
- How many "hit points" are wounds of different shapes worth.
It's linear in that respect. The size of the wound is more important than its shape, and in this case the formula is D2*2, where D is the projectile's diametre in millimetres. Then PF determines the depth of the wound to get its approximate value. I tend to round this for most values, and usually round up. For instance, a 7.62 should have a puncture of ~116, but I gave it 120. As for non-penetrative trauma, it's usually based on their energy (or momentum for bludgeon damage) and the concentration. It's not exact most of the time, but the rounding errors are fairly small.