Just a couple of comments that didn't yet make it to the 'outbox' ;-)
John Wojnaroski wrote:
John writes:
Thinking a bit about the folks playing in the virtual ATC world. Would be nice if FG could be included, might obviate the need for an AI system. Conversely, developing an AI/ATC system is good exercise for the brain muscle and provides a nice alternative for network-challenged machines
;-)
David writes:
Personally, I think both are important in the long term, but since I've started on AI/ATC I'll carry on with that for now. The security implications of coding networking code frighten me - I think my coding style might be a bit, erm, "rough" for that!
Actually, I would not recommend to really invent something new, either -
So in case that we should really need to do networking stuff:
I would rather recommend to use any of the existing networking libraries for that purpose, that way one could mainly concentrate on really developing the *protocol*, there are numerous other projects that do already an extremely good job with the networking stuff, so why waste your time developing logics that others have finished already ...
So, for FlighGear purposes, the "Torque Network Library" ( http://opentnl.org ) looks suitable, as it is cross-platform, GPL and despite from that I have personally done some smaller things using it, so it's really no rocket science - you can do everything in a very abstract way - additionally it is also under active development by a _group_ of developers, which is also the reason why the very same library is used by commercial applications/games.
To be honest, personally I think that developing a protocol in itself is already challenging enough - taking into account that it should be efficient ...one shouldn't waste to much time with the low-level part (i.e. implementation)
But all this may not even become necessary - maybe we can even use the VATSIM/IVAO networks :-)
I just sent a question off, about what kind of variables/data the networks needs to fetch from FlightGear and in what format that data needs to be made available.
I think, that way we can look into FlightGear property tree and determine, what's already there and what isn't yet.
Depending on whether the protocol is natively integrated using plain old C++, or if it gets added by using the I/O protocol XML-based routines, making the data itself available should not really be such a problem, I would say ...
Definitely a lot easier than doing the AI stuff :-)
what we need is the Grand Unified Cosmic Controller Interface (GUCCI) ooooo! (-10 points from Gryffindorf for bad punning)
I sent an email off to the "virtual ATC" folks a few weeks ago (ivao and vatsim) regards hooking in my 747 sim; so far no response. Will be interesting to see if Boris hears back...
What I learned so far:
"Squawkbox" - the windows client for FS 200x simply fetches/puts data using the IPC mechanisms implemented via fsuipc from schiratti.com - so, essentially this software also uses a different set of "common shapes" for all aircraft, these shapes are then displayed for other traffic, within the simulator - so, basically all other traffic is referenced within the coordinate system, and depending on the type of aircraft, the squawkbox application seems to make the simulator display an appropriate 'shape' / with a special livery for each aircraft.
You can find more out about this "common shape library" at:
http://homepages.paradise.net.nz/seang1/csl/
However, this seems to superceeded now - there's a new approach, called:
"Multiplayer Traffic Library (MTL)"
essentially, it might already be sufficient to make other traffic
available within the property tree, with default shapes - adjustable airspeed parameters and other stuff, that way the network protocol
could simply change the variables of the corresponding aircraft
using the prop tree.
I also sent just an inquiry off to the vatsim folks about the currently used VoIP technology, so that we can decide what type of software might be used for FlightGear - depending on the protocols/compression that needs to be supported, there' certainly a lot of open source stuff.
-------- Boris
_______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel 2f585eeea02e2c79d7b1d8c4963bae2d
