On 6/9/06, Jonas Koelker <[EMAIL PROTECTED]> wrote:
Hello, all.
I've been lurking on the list for a while now. I'm really interested in the
project, and can't wait to see a consumerish card come out.
However, I'm in a bit of a peril: I don't know shit about hardware hacking.
Okay, that's not strictly true, I've completed the obligatory 7-week machine
architecture course as part of my comp.sci. major, so I know what a clock edge
*is*, but from there to implementing (even parts of) a graphics board at least
*seems* to be wide and far.
I don't know anything about hardware description languages, and I haven't read
`Graphics Cards: Design and Implementation' (if such a book exists ;)), so I
can't decode how far in the development process you are. A pointer to
resources aimed at people who have only very limited and very rusty x86
assembly as the closest-to-hardware programming skill would be appreciated.
Well, since we're working on OGD1 right now, the discussions have a
distinct bias towards the hardware-design aspects of the project. It
is, after all, a hardware project. But there are myriad other things
to be done that anyone with a clue about comp.sci. can help with.
For instance, you have a damn good idea for a book there. Perhaps you
do or perhaps you don't have the background necessary to write such a
book, but.... say you were to set out to write it. Seriously. What
you'd have to do is start prodding those who work with this for the
information you need. In a way, you might become more of a compiler
and editor (in the human sense) than a writer; you'd take a leadership
role in pushing the book forward.
Lots of people with varied levels of hardware knowledge on this list
have been invaluable in helping to collect information and ideas
together and put it into a form that is digestible to the general
population, as well as helping us hardware engineers keep organized
and focused on our jobs.
Other things we can use help with include web development. Experts
have contributed some great stuff, but you could assist them. There's
research that can be done on all sorts of things (legal, marketing,
technical, etc.). We have a software model of the 3D graphics engine;
some more work needs to be done to make the ideas in it transparent to
other people, particularly those who will write drivers. And we need
to construct a non-profit foundation that handles, among other things,
donations used to buy OGD1 boards for developers who will help us with
the graphics engine work.
Also, since I don't really understand what's going on, I'm having a hard time
discovering if and how I can contribute. I wouldn't mind learning a new
programming language (over the summer break, that is) in order to contribute;
however, I have this nagging feeling that I'm also going to learn the
`hardware programming paradigm' and that it'll be very much different from
most paradigms I know. Any pointers for the regular JACH(*) would be greatly
appreciated.
Yeah, hardware programming is in some ways inside-out compared to
softrware programming. For instance, with software, brevity in source
code often translates into small machine code. But with chips, you
have to be elaborate and detailed. Oh, and going from thinking
serially to thinking in parallel can be tough. (You might want to get
good at pthreads as a stepping stone.)
(*) Just Another C Hacker (... unless $spring :D)
I'm a cheap, poor student, so I'm probably not going to donate money.
However, due to my conviction, it's probably not going to be hard to sell me a
libre-software-friendly graphics card for something like 400$ :)
Well, OGD1 is mostly useful for hardware designers or those wanting to
learn it. It's up to you to decide how much chip design you want to
learn. For everyone else, OGD1 should be seen as a source of funding
for open hardware development and a way to build a new sort of libre
community.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
