[Flightgear-devel] Kudos to the canvas team and to JSBSim (again)

[Renk, T. <tho...@jy...>]

I think it's time to express my appreciation for what the canvas team has created. I have proceeded to modeling the avionics of the Space Shuttle, which has rather information-dense screens with C/W messages displayed on-screen flashing, a hierarchy of both edge-key controlled menus and keyboard command controlled pages and lots of other fun.

http://www.science-and-fiction.org/FG/pics/shuttle_cws01.jpg

http://www.science-and-fiction.org/FG/pics/shuttle_cws02.jpg

I would have originally had no idea how to code that kind of thing.

Using a general framework Richard has developed which in turn uses canvas routines, I find the work of generating all the screens merely tedious (positioning all items on-screen and linking them to properties is not very exciting work) but quite doable. Likewise, I've managed to get trajectory displays and other interesting stuff with a minimum of fuzz - that's a tremendous job you people have done here!


Also, I had another JSBSim moment when I tried an return-to-launch-site abort with the Shuttle for the first time. This probes a completely different regime than the nominal trajectory - basically after ET separation you're far too slow for the altitude where a safe separation is possible and just fall down for a while, reaching vertical speeds of 500 m/s easily - before hitting the dense atmosphere and having to pull out avoiding excessive g-forces. I found that without any tweaks to the aerodynamics, after correcting a few oddities I had in the FCS, I could fly the whole maneuver just as described by the manual. Given that all this is really far from the nominal trajectory, I found that a not quite trivial outcome.

So if you have good input data, JSBSim does really remarkable things.

Cheers,

* Thorsten