Open GraphicsPosted on 2007/09/09 - 00:48
The Finland thing was fun (ended up with a ten, or A+ if you like), I've worked my ass of this holiday to generate enough income for the next school semester, I'm (after a lot of lobbying) allowed to skip a required internship so it's time for the next school semester, or, as I like to say; wicked cool projects. Oh, and I forgot to update this blog. Guess I'm not really having any time to write as often as I'd like. Anyway, this project I'm trying to get pushed through the bureacrazy is mainly motivated by my disliking of another project. If you happen to participate in that project, please don't be offended and all that. Read on.
The project I'm talking about is the Open Graphics Project. This started out as a simple idea to build an open source Graphics Card with some big future ideas and sounded really cool. But, with the third year 'anniversary' coming up next month, all there is is an expensive $2000 fpga board, which was designed without barely any community communication (from my perspective) of which the bugs are identified, not fixed. Meanwhile the mailinglist is working out high speed nanocontrollers with some serious muscle like fullsize 32 bit multipliers, possibly interrupt support, and whatnot, which is supposed to do VGA emulation and DMA transfers in realtime simultaneously. Details about the actual hardware and what's possible is 'underway', but nobody can actually write it since nobody has access to the details of the card. Except for the two people who've seen it.
I partly blaim the creators of the project who are pushing for these design choices. If I can point at the cathedral vs. bazaar analogy, they're trying to build Taipei 101 while half their workcrew is still figuring out how to weld steel. If history has taught us anything, open source projects work best if they start small and get improved on over time. No matter how terrible the first implemenation is, if it's something new and people are interested, it'll get mostly rewritten, redesigned and refined over the years. A 2000 dollar card 3 years after the starting date with wild discussions about nanocontrollers while VGA still isn't working it not 'starting small'.
Now, I'm not usually somebody who complains about other people's efforts. I think it's very bold to try something like this, I just think their approach to the problem could be a lot better. Which is why my new project this semester is going to be to build a small ($100 range) PCI card with a VGA connector and hopefully some unstable VGA support. If Open Graphics is building Taipei 101, I'm building the Tower of Pisa equivalent. It's going to be an idea realised, proof that something like this can be done, and then hopefully the open source community will pick up on it and improve on the original. And maybe I fail and OGP succeeds, who knows. At least the game is on.
[: wacco :]