Re: [Flightgear-devel] Re: Cloud Rendering

2004-11-14 Thread Paul Kahler
Melchior FRANZ wrote:
* Paul Kahler -- Sunday 14 November 2004 15:57:
It's just an integration problem right? ;-)
No. A problem of fixing his code that is in SimGear already. Your are
hereby invited to submit patches.  :-}
I just knew my message would result in my nomination to do the job :-) 
I suggested it to the list because I don't have time for such a big job. 
I might have time to make improvements though if it's already working to 
some extent. After I get my new PC (before year end) I may just start 
poking around in the code. Don't hold your breath though, if someone 
else wants to look at it, don't wait for me.

Thanks,
Paul

___
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d


Re: [Flightgear-devel] Re: Cloud Rendering

2004-11-14 Thread Curtis L. Olson
Paul Kahler wrote:
Melchior FRANZ wrote:
* Paul Kahler -- Sunday 14 November 2004 15:57:
It's just an integration problem right? ;-)

No. A problem of fixing his code that is in SimGear already. Your are
hereby invited to submit patches.  :-}

I just knew my message would result in my nomination to do the job :-) 
I suggested it to the list because I don't have time for such a big 
job. I might have time to make improvements though if it's already 
working to some extent. After I get my new PC (before year end) I may 
just start poking around in the code. Don't hold your breath though, 
if someone else wants to look at it, don't wait for me.

The 3d cloud code really needs to be rewritten entirely to work properly 
with FlightGear.  Mark's original clouds were implimented in his own 
custom scene graph library which does not play nice with FG's scene 
graph library.  Generally, no two scene graph libraries will ever play 
nice together because of something called lazy state evaluation.  
Changing state in opengl (i.e. switching textures, changing render 
parameters, etc.) can be expensive for a variety of reasons, also 
querying the opengl card for the current state of things to check if 
something is set or not can also be expensive ... so scene graph 
libraries like to keep track of all the opengl state themselves, and 
only send changes to the card when something actually needs to get 
changed.  This is usually great for efficiency, but if you have two 
independent systems trying to track and make changes concurrently, you 
end up with a big mess ... kind of like two people trying to drive the 
same car at the same time to completely different destinations ... you 
end up with a big mess, if not an all out fist fight ... at best, you 
don't end up anywhere near you wanted to go.

The current state of things is a lot of bandaids, bubble gum, duct tape, 
and wishful thinking to try to get them to sort of work together ... or 
at least try to keep the 3d cloud code from stepping on the FlightGear 
rendering code too hard.

Regards,
Curt.
--
Curtis Olsonhttp://www.flightgear.org/~curt
HumanFIRST Program  http://www.humanfirst.umn.edu/
FlightGear Project  http://www.flightgear.org
Unique text:2f585eeea02e2c79d7b1d8c4963bae2d
___
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d


RE: [Flightgear-devel] Re: Cloud Rendering

2004-11-14 Thread Norman Vine
Curtis L. Olson writes:
 
 Paul Kahler wrote:
 
  Melchior FRANZ wrote:
 
  * Paul Kahler -- Sunday 14 November 2004 15:57:
 
  It's just an integration problem right? ;-)
 
 
  No. A problem of fixing his code that is in SimGear already. Your are
  hereby invited to submit patches.  :-}
 
 
  I just knew my message would result in my nomination to do the job :-) 
  I suggested it to the list because I don't have time for such a big 
  job. I might have time to make improvements though if it's already 
  working to some extent. After I get my new PC (before year end) I may 
  just start poking around in the code. Don't hold your breath though, 
  if someone else wants to look at it, don't wait for me.
  
 The current state of things is a lot of bandaids, bubble gum, duct tape, 
 and wishful thinking to try to get them to sort of work together ... or 
 at least try to keep the 3d cloud code from stepping on the FlightGear 
 rendering code too hard.

It really isn't all that bad, after all the code is all isolated in one place 
and 
therefore the *expensive* state switching code is only done once per frame.

IMO The biggest problem is that none of us have takem the time to really
understand Mark's scenegraph and since there is no pbuffer support in
PLIB it is rather difficult to translate into a native FGFS style.

Cheers

Norman



___
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d