On Monday, June 10, 2002, at 05:35  am, John Wojnaroski wrote:

> Well, I've been working on a Multi-function Control Display Unit (MCDU)
> which connects to the FMC and
> provides manual control of the nav receivers and other aircraft systems.
> Next major effort is adding ability to
> insert and edit waypoints and do some flight testing of the LNAV and 
> VNAV.

Hmm, this is going to get fiddly. I've got a generic route/flight-plan 
backend sort of working (no flight traversal or waypoint sequencing as 
of yet), which I am planning to extend with airways, SIDs, STARs and so 
on (with editing of course). The long term plan was to build an FMC / 
MCDU / GPS on top of this, but it seems like you're a good way ahead of 
me. And of course this is built into flightGear so that it can access 
the exiting Navaids, fixes and runway data.

> The not-so-good-news --- it runs as a single opengl app (I've got my 
> copy
> running on a laptop) and uses the SimGear library socket functions. (You
> need 3 networked machines, but you do get the *whole enchilada*)
> A full complement of flight deck displays (OpenGC), a multi-function 
> control
> unit (MCDU), a flight dynamics model
> and scenery(FlightGear).

I was assuming long-term I'd setup an /fms node in the property tree, 
with a flightplan [fp?] child containing the waypoints of the active 
route  (with a few other fp slots for secondary and temporary routes, of 
course), and that would be enough data to generate a reasonably complete 
and accurate Boeing / Airbus style NAV display, or the list / visual 
display of any smaller GPS unit (KLN-89b, for example). However, this is 
some way off (I want to get the flight-plans hooked up to the existing 
auto-pilot code, which isn't really designed to look like a 'big jet' 
model of course, though it seems you already have code *somewhere* that 
handles LNAv and VNAV functions... hmm.

I'm not sure the best way to proceed. I really think the auto-pilot code 
belongs in flight-gear, but the goal of having the FMC / MCDU run 
separately is valuable. There's a lot of navigation  data the FMC/MCDU 
needs that is simulator specific though : does your code have it's own 
copy or what? Also, people who just want an in-cockpit GPS unit 
definitely don't want a separate machine to run it on.

One solution I've contemplated (but which I suspect is unworkable) is 
something like an NFS-mount for the property tree: i.e run the FMC/MCDU 
as a  property 'server', and have every node it exports be located under 
/fms in the global tree. This would probably get inefficient and / or 
too complicated in short order though.

Anyway, I'd like to work out a plan of attack to avoid too much 
duplication. I suspect some might be inevitable, since there's slight 
different goals. (and, alas, I only have one computer, so I'm kind of 
anti the whole 3 computers thing :-)

H&H
James

--
Anyone who considers arithmetical methods of producing random digits is, 
of course, in a state of sin. --  John Von Neumann, 1951




_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to