Originally posted by Vermillion:
Do you have a background in Physics? Mechanical or Aeronautical Engineering? Math major? Or any field of Engineering at all? Worked in a wind tunnel or other type of flight testing laboratory?
I have a math minor, does that count?
this part reflects in no way how things are modelled in either game. Just a basic thought about real time calculations vs table look ups. Let's assume you've got the same input variables in both cases. In the real time equation, the variables are inserted into the equation, and the algorithm kicks out a value. The same value, every time, given the same inputs. If it doesn't, something is screwed up (or there's a random number generated somewhere in the calculations.) basic example is X + 2. I put in x = 1, I get 3, no matter how many times I do it. Now, if you use a table look up method, you grab all the input variables and using them, you pull a value out of the table (could be using a relational database with a SQL like query system, doesn't matter for this discussion.) Point is, given the same inputs, the same result will be achieved every single time the lookup is done (barring a random number generator.) Everyone hopefully is agreeing with this...
Now, let's assume that you're using these methods to calculate the lift generated by an airfoil. I never took fluid mechanics, but I'm pretty sure there's a standard equation to calculate airflow over a surface, from there you determine the difference in pressure over the top of the wing and the bottom of the wing, and from that you get the lift. Some (yes, some, I don't know 'em all) of the variables you'll need are airspeed of the foil, the density of the air (this would come from altitude, humidity, temp, etc) and the shape of the foil. These inputs are the same regardless if it's going into the real time calculation OR the table look up. The difference is the lookup table values have already been calculated. Therefore you know in advance what the lift for THAT foil is at THAT speed, THAT altitude, THAT air density, etc.; all you do is pull out the value. The equation just chugs through the variables and spits out the answer. The answers should be nearly identical (difference coming from the fact that you'd have to round the table lookups at some point, the real time should be able to take all floating point numbers.)
The point here is that a real time calculation doesn't give much of an advantage over a table look up (if the table numbers were generated with the same algorithm used by the real time calculation.) It can give a finer degree of control due to not having to round data before being fed into the method. However, if the algorithm is intensive, you'll lose efficiency. The table look up will be faster. In other words, saying X is more fluid because it uses real time calculations vs AH look-up tables is a little off base, IMO. SW is right, we don't have any idea how either sim really is deriving its values. For all we know, given the same set of inputs, both sims might spit out the same outputs!
The difference in "feel" for you might be other aspects of the software. Perhaps it's the graphical API calls being used. It might be the fact that AH is designed to be multiplayer, so the planes feel a little jerky, because the game has to keep track of the other planes as well as providing flight data for yours. More factors than just calculating the 4 forces acting on a plane (thrust, lift, drag, gravity) go into how the flight model "feels" to each individual.
The second point I'd like to make is this. Let's assume that X-plane's FM IS better for the moment, and everyone agreed on it. Should HT and crew tear apart their code to make it fly more like X-plane? Before they could even do that, they'd need to get the algorithms from the X-plane people, if X-plane would even let them have it. If not, then HT and crew would have to reverse engineer it (perhaps even come up with it on their own), then write up the design docs, do some initial verification and validation before even setting out to code it. Then they'd have to code it, test it, recode, test, etc etc. Then they'd have to integrate it into the rest of the game code, which could be a nightmare. Basically, it'd take anywhere from 3 months (yeah right) to a year just to do it. Personally, I'd rather see some new planes, railroads, strat changes, etc in that time frame than a re-done flight model that "feels" a little better.
