One of the key points is also thinking about what is COMMERCIALLY viable and worthwhile. I know everyone hates the thought - but if we don't make money, we don't stay in business and RO dies.
There is, therefore, the question of "is it worth it?"
Our view(s) run something like this on the topic:
1. We want the game to be as realistic as possible, while still being playable.
2. There is very limited return (money) in building complex stuff that only 5 people will even notice, never mind understand.
3. We also have to load-balance: there is a ton of things it might be fun to implement, but it all adds up to more load on servers that are already stretched.
So - ballistics.
We don't model internal ballistics - realistic, but an utter waste of time. I'm sure some people might like us to, but we simply aren't going to bother.
We model the major characteristics of external ballistics, including speed, air density/temp and round shape (as characterised by the ballistic coefficent) and some other odds and sods that John can remember. We don't bother with the odd first few milliseconds before the bullet fully stabilises in flight, nor do we model the effects as the round starts to destabilise again down-range. Firstly, these effects are really only of interest at very long ranges, or to specialists. 99.9% of the playing public (and shooters) will never even notice the millimetric differences involved. Some way off in the future, we may add some additional detail on large-sized rounds at long ranges - but we need ranges of well over 1,000 metres for those to become even vaguely visible, even with an 88mm round. And we probably aren't going to get on to such ranges until we get on to UT3. Another consideration is that we are playing this on screen, not with the naked eye - there is more than enough inaccuracy in a single pixel at that range to make the exercise pretty pointless. We'll see in the future
And terminal ballistics... we provide a basic level of terminal ballistics modelling for main gun rounds - more than most other games have ever done. We don't bother with anything clever on small arms rounds. Fun as it might be to model the tumbling of a round when it hits "meat", the glancing and shattering, it simply isn't worth the CPU cycles. For game purposes we are only interested to see if the hit is disabling or not. We have stated many times that we will continue to improve on the terminal ballistics/damage model for main gun rounds. It will take time - we have to find ways to do it without crippling the servers which, frankly, is the biggest challenge. We have to do it bit-by-bit, test it out, blah blah. We're not a big development house with $5m and 2 years to play with.
So - if there are aspects that people would like to model - go do it! You'll find plenty of the source code around to play with (not for the engine - Epic obviously don't allow it!). Just because we didn't encourage people to mod-the-mod years ago has nothing to do with today's situation. In case people hadn't noticed, we've got Valve to list RO mods in Steam - we WANT people to mod the game.
And penetration of objects: of course it can be done. But, as the engine doesn't have it built in, it is a complex job to code WELL. As John says, it is also potentially a big load on server CPUs. Again - if you think it can be done easily, then go do it. If it is easy, light load on CPU and really works well in the game, we'd probably buy it off you
Summary: it is very easy to talk of building many of these things in - but we have to live with the commercial realities. Our internal aim is to follow a "90-10" rule - make the game "90% realistic" for 10% of the cost (and this hasn't been a small 10% over the last couple of years!) - the remaining 10% would cost a further 90% and we'd go broke!
Don't get us wrong: some of these things would be fun to do - just that we don't think they'd add enough to the game to pay back. Therefore they are actually best done by those who don't NEED to make it pay back - go make the mod
Edit: simple definition of terms:
1. Internal ballistics - the performance of the round in the barrel
2. External ballistics - the performance of the round from when it leaves the barrel until impact
3. Terminal ballistics - the performance of the round after it impacts whatever it hits; this is the charming catch-all term for damage modelling, including how a round rips up bits of people...