Hello,

I'm a student Computer Science who unfortunately misses the kowledge and
free time to really help with the Open Grapics project.

In reply to dsp Mind's email, last year we had a project where we were
introduced to hardware programming in VHDL, so exactly the group a)
users. However, I don't think that OGD is that interesting for that
group, because it's too expensive ($250 or more?). We used Spartan-3
Starter Kit Xilinx Development Boards (http://tinyurl.com/4z7yk), and
those are only $100. Of course OGD has much more ram, features and better
FPGA, but even the starter kit was way overkill for what we did (playing
with LEDs, switches, buttons and stepper motor, making a crude elevator
simulator). So I would concentrate very much on group c) and d), the more
serious users.

As for the reasons why there aren't many successful open source h/w
projects:

Reason a) Is not totally true as stated above, but it is for hardware
which can be put in a computer (e.g. PCI cards).

I think the real reasons are that there is just much less expertice in
hardware programming and the nature of hardware.

Timothy's hate of Wishbone illustrates the first point. Though it works,
it apparently isn't good enough. And when reading this mailinglist it
becomes clear that designing hardware isn't as easy as software
programming, a lot of experience seems to be needed to achieve something
at all, while in software it's more like everyone can make whatever is
wished, and it will always work, only how stable and ugly the code is
differs.

Software can be easily copied, hardware can't. So sharing hardware
designs in a useful way is much harder in practice. But FPGAs can change
that.

It would be great if there was a cheap FPGA add-in card which can be
easily (re-)programmed at runtime, something like OGD, but less
expensive. Then people could write sort of FPGA code modules for those
cards which can be used as sort of libraries by normal programs.
Something like the early 3D cards, but then for programmable hardware.
Progams could use those cards to accelerate certain things, like
compression, encryption, de/encoding, or whatever else is the
computational bottleneck in the application. If people could get such
card easy and cheap enough (starting price $25-50) and it becomes
popular, before you know there'll be a lot of people much more interested
in hardware programming.

The reason OGD can't be that card is because it's own hardware demands
are too high to be cheap, and it is a videocard, so in general it's used
all the time and can't be easily used for different things.

Greetings,

Indan

P.S.
Everything alright with the opengraphics.org Wiki? All content seems to
be lost, and earlier today the site was down.


_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to