I've been reading some great books on software rasterization and I believe this stuff is very important to learn for graphics programmers (and there's no way around it with the GBA). With such a high interest in building software-based engines, would you guys like a website (or blog) targetted for the PC and the GBA? My idea right now is to teach efficient algorithms on how to solve certain tasks. What are your thoughts? Any other ideas are welcomed too.
Phil
I don't know about GBA, but the PC parts would be cool
I have no idea what you just said :P.
But as long as rasterization means free money, then yes :).
Yes our engines are coded in software, but we use hardware accelearation to rasterize stuff, i think you are talking about implemetation without hardware acceleration, which is also important to learn.
Yeah, I'm talking about the discussion of building software-based 2d/3d engines. The major benefit I find is we're better graphics programmers if we understand what's going on behind the hardware acceleration. With the GBA not having hardware acceleration, the same thing I teach on the PC is instantly the same for working with the GBA. I think the PC will be the major focus, though.
The software rasterization days are gone.....thankfully.
It is good information to know and understand, but when DX and GL both do it for you, I see no point in writing and engine that uses software.
Yeah, it'd be pointless to write your own software engine when hardware is present. But I believe we, as graphics programmers, aren't as great as we could be until we understand what's going on with the hardware to draw objects to the screen. If it's not interesting to the fellow programmer, is he a good graphics programmer to the company? Even I don't have an answer to that, but I believe he'll have an edge if he does, and he'd be more employable than the average joe. To be fair, in the real world, you're employeed to get a job done. But I guess I just find it cool that I don't have to depend on the video card to get some extra work done. I guess that's what drives me into creating this informative site. I'm also thinking of adding tutorials on optimization theory with C++ using DirectX. Thanks for the thoughts everyone.
You don't have to code a software rasterizer to understand what it does behind the scenes or how it works.
I do think professional graphics programmers should understand how software rasterization works, but I don't think it is necessary to have coded one yourself.
In very certain conditions coding a software rasterizer isn't *that* hard. Do it in a high level language, and use hardware acceleration but manually project/clip and fill the polygons, alter the color buffer, but then let your API and video card to the work of sending it to the screen.
Doing a software rasterizer at a *very* low level, with as little outside help from an API, is a pain (just doing something 'trivial' like sending a color buffer to the screen is by no means particularly easy).