October 30, 2002

Still working on the DX8 port... I've been able to convert all dialog boxes (and visual components, obviously), all windows (with fade-in effects, eh eh), and today I finished the main menu. Well, nearly finished. I'm still having a few minor issues with the highlighted text and shadows, and I haven't found a workaround so far. But I probably will, uh. So, basically, most of the game is now powered by DX8, although I'm not so happy about it. I mean, I was expecting a jump in performance, and there's actually a 10-20% drop instead. It's even worse during the game selector, where the 3D logo animation only reaches 170 FPS when it used to be at 230 FPS with the DX7 version. Okay, at this level, there's no real point in gaining performance or not... I know. But I'm still thinking about the older video cards that have trouble running the game at full speed (sob). I believe DX8 will only be a transition for me. From now on, I'll be able to decide whether I want to switch (back) to OpenGL, or give DX9 a try. I've been enrolled in the DX9 beta program today, and am looking into the 200mb SDK. It's absolutely useless to me for now (there are no Delphi headers provided, and the people I know who usually do the header conversion aren't on the DX9 beta program, so I'm stuck with the C++ headers which aren't of much interest to me), but at least I can start learning what DX9 could bring me. I was first interested in the library because it re-implemented DX7's "GetDC" instruction [allowing GDI operations on DirectX surfaces] which Microsoft seemed to have given up in DX8. Only, I've now made a workaround by having a GDI surface in memory and copying its contents manually into a DX surface whenever the GDI stuff is modified. It's a rather CPU-intensive routine which manages to noticeably slow down a Pentium III 450mhz, unfortunately, but not enough to bother me. On my new Pentium IV 2.8ghz, the routine is just as fast as what I used to get with the GetDC method. One of the advantages of my manual method is that I could port the game to OpenGL without many changes needed. So I'm sticking with it for now... Overall, I haven't been very creative this week. The only changes I made to the game are cosmetic, and these are very small graphic details that nobody but me will notice. Ah, never mind. At least I'm currently reworking a lot of code to make it cleaner, delegate low-level tasks to subroutines, and make the code generally easier to understand and to port to another API -- hopefully. I'm still a long way from achieving a professional-looking source code, but it's still better than what I had last week. And I'm not planning to stop there. Maybe I could implement a dual-engine to enable both OpenGL and DirectX support. Ahhh, silly me. I know it would never work... I'm too lazy to do this kind of thing

Anyway. It's bedtime for me now (I know, you're relieved !). I'm not too sleepy though. I don't want to count sheep. What about DirectX versions. DirectX 2, DirectX 3, DirectX 5... Brr... Zzzz..... PS : I know that "only" 86% of you don't seem to have problems with OpenGL, but I have one question... Why, out of the remaining 46 voters, has no one taken the time to send an e-mail explaining the crash ? I'd specifically asked you to write... Hu. Maybe you're just teasing me and you didn't even try version 10.21 ;o)