I have been building a (highly accurate) flight simulator.
Basically, I'm taking a geometric approach to the flight dynamics model (similar to how x-plane works, except they're using... older techniques/technology. I'm using much more advanced stuff [working on seeing if I can simulate the airflow for the entire craft, as opposed to doing it by sections and then integrating those together]). This is in contrast to things like MS Flight simulator and FlightGear (though, flightgear does have a poorly documented and rather inaccurate geometry based fdm - but most people use the table-based one, which is far more accurate than the geometry-based fdm they have implemented), which use lookup tables to guess how an aircraft would perform.
The problem I actually originally set out to solve was that xfoil and xflr5 suck to use (importing/exporting plane data is... either you can't, or you shouldn't), and fuck paying for the more expensive design testers. (this is why I have the focus on accuracy - if this was just going to be a simple game, I'd have spent far more time making it look pretty) However, I figured that I could also make testing be more fun by adding an interactive mode (i.e. I want to be able to do hardware-in-the-loop type stuff, as well as just manually flying), and at that point, it just is a scriptable flight simulator.
I'm still working on the flight dynamics model, been teaching myself fluid dynamics so that I somewhat understand what all is going on there (as much as anyone who hasn't spent years studying this can understand...), and I've been working on writing code to run on the gpu (yay, opencl) in order to do this. It's been fun.
That sounds great. As someone who has a lot of interest in aviation and stuff that's related to aviation, the lack of flight simulators on the market is bothering. I tried almost every simulation that's out there, but none of them fit to my needs (good, simple UI with extremely realistic controls. that's not too much to ask, is it?). Hopefully you'll make something that's more appealing than other products. Speaking of which, is there a website of sorts to follow the updates of your work?
Thank you for the interest! Right now, x-plane basically is the only player in the flight sim market (MS has officially stated they've left - though I doubt it's for good, FlightGear is the only other flight sim worth mentioning, and it's nowhere near as good/accurate as what I need), which, as you (indirectly) pointed out, is troublesome - the market is ripe for disruption. I'm aiming to at least provide healthy competition to increase the quality of all flight simulators (I'm also applying to get into this next YC batch... hopefully I'll get an interview. I doubt I'll get in, because lack of a cofounder).
I actually just wrote a script to generate an atom feed for this (I mean, I have some updates which are on my main index page, but, that's not a way to easily enable people to follow me...), so you can follow my little self-generated atom feed at http://younata.com/feed.atom right now, it's just a single post that links to my above comment, but I'll try to write at least once a week.
By the way, the datestamps on that feed are in UTC, which is localtime for my server.
Feel free to send me an email so I can notify you when I have a prototype. My email is listed in my profile.
Yes, it's a shame. I still think FSX beats the pants off X-plane. I just find that X-plane isn't as realistic as FSX (even though it should be in theory). Anyway, good luck - it's a massive job :)
very very cool! At some point i want to dig into how to write physics simulations, in the mean time i'm slowly working on trying to make better numerical computing tools, which is sort of a precondition to playing with simulation.
Oh, man. My previous experience in building simulators is a spaceflight simulator I wrote at the beginning of the year, because space, life is easy and I didn't have to really write hard numerical computing code. This is easily the most challenging thing I've ever written (but it's also easily the most fun thing I've ever written).
Good look getting into writing physics simulations!
Basically, I'm taking a geometric approach to the flight dynamics model (similar to how x-plane works, except they're using... older techniques/technology. I'm using much more advanced stuff [working on seeing if I can simulate the airflow for the entire craft, as opposed to doing it by sections and then integrating those together]). This is in contrast to things like MS Flight simulator and FlightGear (though, flightgear does have a poorly documented and rather inaccurate geometry based fdm - but most people use the table-based one, which is far more accurate than the geometry-based fdm they have implemented), which use lookup tables to guess how an aircraft would perform.
The problem I actually originally set out to solve was that xfoil and xflr5 suck to use (importing/exporting plane data is... either you can't, or you shouldn't), and fuck paying for the more expensive design testers. (this is why I have the focus on accuracy - if this was just going to be a simple game, I'd have spent far more time making it look pretty) However, I figured that I could also make testing be more fun by adding an interactive mode (i.e. I want to be able to do hardware-in-the-loop type stuff, as well as just manually flying), and at that point, it just is a scriptable flight simulator.
I'm still working on the flight dynamics model, been teaching myself fluid dynamics so that I somewhat understand what all is going on there (as much as anyone who hasn't spent years studying this can understand...), and I've been working on writing code to run on the gpu (yay, opencl) in order to do this. It's been fun.