What's the catch?

I’ve been reading through the Panda manual and am getting really excited about using this engine.

For the last several years I’ve developed my own 3D game engines in Java, but am getting tired of all the politics surrounding Sun/Oracle right now. I’m still looking for something cross platform though (Windows, Mac, Linux), and that doesn’t require retooling, recompiling, etc. for each OS.

I have almost no experience with Python though, and am not sure how difficult a transition from Java may be.

SO… with all that mind… what’s the catch? The Panda packaging tools are really exciting for my cross-platform goals… but do they really work? Is Panda going to turn out to be really slow? Am I going to have problems running my games on those icky Intel 945 graphic chips that are installed everywhere? Will I need to ditch Blender and buy a multi-thousand dollar modeling tool so I can import my art?

Any advice or foresight would be greatly appreciated before I jump in!

I’m pretty sure that you can use Blender. If you want to check whether it will run or not, you could try one of the Disney games(ToonTown, Pirates of the Caribbean online) with a free account.

Python is probably the easiest language of all to learn (it’s almost like pseudo code), so you shouldn’t have any difficulties getting up to speed with it. Especially having experience in programming in general. Just read here: docs.python.org/tutorial/index.html and you should be good to go. Once you’re done, move on to the code samples. You shouldn’t have too much trouble getting used to Python.

There’s no catch :wink:. At least I’m yet to bump into any. Of course you won’t get everything Unity or UE3 would give you (like WYSIWYG editors, shader editors and whatnot), but you will get one of the best documented, easiest to use and probably the best supported OpenSource engine available.

No, as long as you RTFM and optimize your code and graphics. Panda isn’t any slower than any other engine, and it supports all important optimization methods, plus it’s equipped with some nice debugging and profiling tools that can really save your ass.

It is by itself written in C++, so it’s not a victim to performance issues high level languages suffer from. As long as you keep all cpu intensive stuff off Python (you can write it in Python and, after identifying bottlenecks, rewrite a specific method to C++) you should not notice any significant slowness with Panda compared to .

That depends on your graphics and level design more than the engine, you know :wink:. If you make Gears of War-level graphics than probably not. If you keep it down to earth – sure.

Panda itself scales very well and it can go from Half Life 1 grade graphics (or even lower) with vertex lighting up to full dynamic lighting, dynamic shadows and a whole lot of shaders on a big, big level.

The Chicken exporter will do the job of exporting from Blender for you. It’s really good, supports almost everything you need except for lights, but it’s not difficult to work around this limitation using empties and tags.

It even supports setting up collisions, at least for the Panda’s internal collision detection system (doesn’t work for ODE).

Have fun using Panda.

oldest/slowest gpu i ever run panda on (except for software-rendering) was a Savage/MX-MV AGP video chipset with 8Mb memory.
Of course all the hardware-limitations imposed by that GPU apply, but panda ran very fine on it and it was even possible to produce somewhat nice graphics. well-made textures are a must-have on low end hardware.
so dont worry. those intel chips are totaly fine :slight_smile: