Lumbering Dream Interactive

Small step for humankind, big step for man: I've finished my first proper game, Pitchblackoids! By first I mean first written in C++, first on a 32bit platform, first not in BASIC, first in a decade. By proper, I just mean that it has a front-end, it has sound effects and background "music", most of the basic facilities (levels and progression, basic AI, pause, highscore save/load, loads of particles and some other effects)... Click for the download links and a quick postmortem.

Read more »

The gamepad support is now complete, and is very sleek. I've used the same approach as with the keyboard input and mouse support: avoiding event handling in favor of a full update on keystates and gamepad axis/button/ball/hat states. Obviously, XRhodes being just a wrapper, the client has access to all the SDL funcionality - if you don't query the keys, the mouse, and don't instantiate and query XR::Joysticks, there's little to no significant performance implication (apart from couple of bytes memory usage).

Read more »

There's been a lot going on recently (never mind posting blogs), here's a quick summary:

• the game controller support for XRhodes is getting done
• not completely unrelated to the above (but not restricted to either), the Input subsystem is getting a revamp
• XRhodes is going Unicode in the near future.
• there are two games in the pipeline (one basically finished and only waiting for the gamepad support), NOT including Diamond Hunter 6000. These are entries for TIGSource's Commonplace Book competition.
• Diamond Hunter 6000 isn't abandoned either.

Arguably the most interactive XRhodes tech demo is now available for your downloading pleasure!

Read more »

Yeah... as I figured, 1000+ particles on the screen at a time is way more than enough. And Grab eats more FPS than the code. (The values are based on a 2.4GHz Intel Core Duo iMac & OSX 10.5.4.)

Just about two weeks outside of the development of XRhodes, and it was a real pain to get back into it. But I managed to. First, I've quickly refactored the Shape system. Again. A lot of magic had gone through the window, but it's the simplest and probably the most efficient now. It's only that it incurred more coding than the previous templatized versions. Oh well.

More importantly, I've brought the Particle System up to scratch. Nothing to be published yet -- the performance implications are quite shocking, or so I guess: with 1024 textured particles the 1500 FPS performance drops below 300 FPS without sublime collision calculations for particles, yet. On the other hand there might be no circumstance where 1000 particles need to be present on the screen simultaneously (and 300FPS is still nifty on its own). So we will see.