Re: [certi-dev] CERTI billard test / http_proxy

[Martin Spott]

Erik, Petr,

Eric Noulard wrote:

> If you find it useful, tell me when you are satisfied with your CERTI CVS tree
> and I'll tag that one in order to be able to base the release on this 
> particular
> source tree state.

Ok, I think today I'm satisfied - look:

  http://foxtrot.mgras.net/bitmap/FGFS/KDFW-HLA_ASN-c172_01.png

....  built from todays FlightGear/CERTI/VirtualAir as of approx. one
hour before.

My hopes with CERTI/HLA as an infrastructure for multi-instance
simulation (aka 'MultiPlayer' ;-)  is to have proper synchronization
among the involved instances on a local network, thus to reduce
latency.
FlightGear's native MultiPlayer protocol, in contrast, aims at
delivering a somehow 'reasonable' result even over long-distance
network connections with unknown latency. The downside of this approach
is that you're getting approx. 1/5 to 1/4 seconds of latency even when
you're running two instances on the same machine directly against each
other, without any MultiPlayer relay involved. This is a little bit
obstructive for example if you're trying to do calculations of the
force on the rope while performing glider towing ....

Now that I know how to use it in general, I'll be going to perform a
few tests during the next days or so. There are still a few missing
features with the "VirtualAir" interface, but I consider these to be of
minor priority (simulation of flaps position, livery selection and the
such). Most important to me that it's working in general - and this is
really great !!

Best regards,

        Martin.
-- 
 Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------