next »

[trabitboy]

quits after a few seconds,
which is a shame cause it feels very fun   
feels like a winner if you can get a stable release out 

[Pishtaco]

I've rehosted it, with avbin included: http://www.willhostforfood.com/access.php?fileid=58921 . Just running flight.exe should hopefully work.

I think the trouble with rolling is that I've included some automatic rudder when you bank, because I read something once about co-ordinated turns. I will change it so that at higher speeds the elevator and ailerons have reduced effects and the rudder doesn't do anything; maybe that will take care of it.

[fiber optic asparagus]

this is one of my favorite games already

[trabitboy]

works perfectly now 

[agj]

Amazing aesthetics, nice feel, though I have a hard time getting the plane to go anywhere but down...

[Pishtaco]

Amazing aesthetics, nice feel, though I have a hard time getting the plane to go anywhere but down...

Try just flying straight, without banking to either side, and experimenting with different throttle and elevator settings. The plane should be pretty stable, particularly as I cheated and put in some forces to automatically cancel out small amounts of banking.

I just spent a night trying to come to grips with using Numpy for rotating and translating a couple of hundred flying triangles. It turns out that numpy can't do lots of linear algebra operations in parallel (which is what I was hoping to use it for), and in fact that it is crap at linear algebra in general. It's running slower than the straight python version -- although this would probably change if I were transforming more complicated shapes, or if I knew what I was doing. There is some kind of fortran library accessible through scipy that is better at matrix multiplication; but if I import it, there is a conflict with pyglet's sound library; and to get it to work well I am going to have to fiddle with memory allocation and suchlike, and might as well just learn C.

[ras]

I really like the visuals and the feel you get from them. Keep up the good work.

[muku]

I just tried to try this on a Vista laptop with a crappy integrated Intel graphics chipset, and this is what I get:

Traceback (most recent call last):
  File "<string>", line 9, in <module>
  File "C:\flight\buildflight\out1.pyz/pyglet.window.win32", line 423, in __init__
  File "C:\flight\buildflight\out1.pyz/pyglet.window", line 644, in __init__
  File "C:\flight\buildflight\out1.pyz/pyglet.window", line 290, in get_best_config
pyglet.window.NoSuchConfigException

 

[Pishtaco]

Here's a new prototype: http://www.willhostforfood.com/access.php?fileid=59583
I see that the executable size has grown a lot. I'm not sure why; probably it's some new python library it is including. Muku: I hope this one works for you. I think the problem before it that I was tring to force anti-aliasing.

Space fires the guns (which are anemic at the moment) and control goes to gunsight view. You can press m to spawn some flying black triangles. These follow you around, but don't do much else right now.

A screenshot (already posted in the art subforum):

[Fuzz]

It's looking nice, although I would rather play it in fullscreen, especially with the problems that come with windowed mouse control.

[nihilocrat]

I like the new model! I wonder what it will play like once you get enemies / prey / whatever in...

[Pishtaco]

I've found a python particle engine that I would like to use, but it only works with Python 2.6. Whereas the physics engine I'm using only works with Python 2.5 . And the author of pyglet has some kind of philosophical objection to getting things to run at lower resolutions in fullscreen.

[nihilocrat]

Are you referring to Lepton? I'm sure you can install it from source, but on Windows that's probably gonna be harder if it's not a pure Python library. I'm betting the dependencies are really low though, so you might be able to build a lepton DLL and then just "setup.py install" to install it using your 2.5 install.

Interesting, I didn't know that about Alex. Honestly, I haven't tried getting any of my pyglet games in fullscreen because developing in fullscreen is a pain... you have to wait for your computer to switch video modes and all...

[Pishtaco]

Are you referring to Lepton? I'm sure you can install it from source, but on Windows that's probably gonna be harder if it's not a pure Python library. I'm betting the dependencies are really low though, so you might be able to build a lepton DLL and then just "setup.py install" to install it using your 2.5 install.

Interesting, I didn't know that about Alex. Honestly, I haven't tried getting any of my pyglet games in fullscreen because developing in fullscreen is a pain... you have to wait for your computer to switch video modes and all...

Yes, Lepton. I had a go at compiling it, but it seems to want Visual C++ 2003, or something like that. I tried compiling PyOde (the physics engine) for Python 2.6 instead, but it's all looking rather complicated so I've given up for now.

It seems that he thinks that the fullscreen stuff is very clunky, especially for something that's meant to be crossplatform; and that it shouldn't be necessary for games written nowadays. I can see the first part of this, but his workaround (drawing to a texture and then blitting a scaled up version to the screen) seems slow and isn't working well for me at the moment; something weird is happening with the alpha values. Anyway, sorry to be petty; kudos to the guy for writing pyglet and making it available.

[nihilocrat]

Hehe, I didn't mean to call you out for being petty or anything, I guess what I wrote ("I didn't know that about Alex") looks funny, like it was a personal attack or something. I agree that it's bogus that he hasn't got a good solution for fullscreen and doesn't seem worried about it. It might be the case where fullscreen is much more clear-cut in Linux/Mac but gets more complicated in Windows.

It's stuff like this why I keep to bigger engines like Ogre3D or Panda3D for my 3d stuff... it's easier to do everything, though honestly they have their quirks where the developers decided to completely leave out or half-implement something you feel is critical. Downside is, the applications are beefier, slower to boot up, take up more disk space, etc.. In this day and age my game (weighs in at 78 mb of RAM fully loaded, 63 meg on the disk) isn't exactly breaking the scale, but is a complete porker when you compare it to 2d indie games that are 10 meg or less. I don't want to get in a tangent, but my experience with Pyweek suggests there are people out there who are apparently disgusted by compo games of this size.