Re: [pygame] Re: SDL 1.3 blit speed

[Brian Fisher <brian@xxxxxxxxxxxxxxxxxxx>]

You should continue programming with whatever system allows you to continue programming with your game. Productivity of making your game should be your primary metric here at this time, far above all other concerns.

If pygame has some obstacle now that's preventing you from moving forward with your game, and you know of an alternative that would be more productive in getting you to your goals, you should go ahead and switch now. If their is no obstacle to moving forward though, and pygame is the most efficient way for *you* at this time, then you should not switch or worry about future concerns now. Usually future concerns get easier to resolve as they get closer to the present.

To explain a little more why I say this - the cost of reworking a playable game to use some different backend for it's rendering/input/audio/etc. is usually very very small compared to the cost of making and finishing a game.Also, swapping out backends is boring and somewhat tedious work, which is easy to be motivated to do for a mostly complete game you really like, and is very hard to find the energy to do for a game that is early on in development, or still much in flux. Finally, it's always possible to have a "pygame compatibility layer" for anything you'd replace pygame with - meaning you could make a module named pygame that looked like pygame but ran something else underneath - so you could potentially do most of the work to change your game off pygame that way (with some small adjustment for the stuff where that isn't sensible), and the size of that work is largely a function of pygame itself, not of your game, so the work to switch doesn't get any bigger by putting a planned switch off.

so I'd turn the question on you - is pygame with it's SDL roots working for you?

---
P.S. As far as your performance question of SDL 1.3 being 3x faster - it depends on your situation, and the answer may not even matter to you. SDL 1.3 would have much better hardware acceleration backing, which could easily be 3x faster for rendering/drawing, but it would most likely be different enough in terms of behavior where you'd have to do some reworking to your code to see that benefit. If you were using OpenGL with pygame now, or using pygame's HW accelerated blits exclusively, you may not see that rendering benefit. Finally, if your bottlenecks aren't the stuff that SDL 1.3 would make faster, it's improvements wouldn't matter to your game anyways.

On Mon, Jul 11, 2011 at 4:41 PM, Brian Brown <brobab@xxxxxxxxx> wrote:

I've been working solo on a large game for the past two years, and I need to know if I should continue programming with Pygame and SDL.

On Mon, Jul 11, 2011 at 4:32 PM, Brian Brown <brobab@xxxxxxxxx> wrote:

Hi pygame users, will SDL 1.3 graphics be at least 3x faster than the current SDL graphics?

Thanks.