Subject: RE: [linux-audio-dev] open-source like hardware
From: Pieter (pieter.palmers_AT_student.kuleuven.ac.be)
Date: Mon Jan 14 2002 - 12:12:58 EET
> anyway, i personally think you're mad. software is truly amenable to
> "community development" because the costs involved in getting started
> are relatively small. this will never be true for hardware unless all
> you are doing is hobbyist-type stuff that is never really there for
> anyone else. i very, very much doubt that you could build a device
> with the capabilities of the hammerfall for less than the cost of a
> hammerfall.
>
My common sense says that you're right. But then again, I'm still too young
to get restrained by common sense.
PS: The most costly affair in a hammerfall-like design is the
software/firmware.
If you look at the hardware (board-level) design, without the firmware, it's
not
that special or time consuming. The real challenge is the firmware &
software.
The above doesn't really apply to the hammerfall, as it uses an FPGA, and
altough
FPGA's are programmable, the programming is more like HW design. The result
is that
the hammerfall board-level hardware is even more simple than a DSP based
board, but
the FPGA programming (configuration would be a better word) is a lot harder
than
programming a DSP.
I agree on the fact that software design is very different from hardware
design.
I also agree that the open-source model has little chance in working for
hardware.
But I never sayd we'd have to copy the open-source software model. I think
it's possible
to develop a model that addresses the specific hardware needs.
The main reasons GPL h/w design, imho, won't work are:
1) Hardware doesn't evolve (much), once it's designed not much changes. So
if you
make a great design, and make it 'GPL' chances are big that some company
will adapt
your design and make lot's of money with it. SW design differs in the fact
that developement
never stops, and people don't like to spend money on software. Spending
money on HW
is the most natural thing. Leading me to the next point:
2) Hardware design is more profitable than software design. Entering the
hardware
market requires more resources than entering the software market (where a PC
will do).
So the competition isn't as hard. And as I stated before, people are
prepared to pay
for hardware.
3) Almost anyone can type "./configure;make;make install". Not everybody can
make or
assemble a PCB (not that I can make 4 layer boards packed with SMD devices).
So spreading
PCB layouts & schematics don't make much sense. On the contrary, for reasons
stated in 1
I would be rather reluctant of doing that. But keeping it in-house means one
will also
have to produce them in-house... etc...
4) I think one can come up with a lot more reasons why it might not work,
and a lot of
reasons why it might work.
The bottom line is:
I want to do this. I don't care if I can't get it to work... the expierence
remains. (I am EE studen so it's welcome)
I am perpared to fund it to my capabilities.
I am not prepared to get ripped off by a company stealing a design.
My goal is to create a quality soundcard fullfilling the Linux community
needs. (I allways set very high goals)
This reply is too long. I stop typing here.
Pieter
----------------------------------------------------------------------
pieter.palmers_AT_esat.kuleuven.ac.be
Currently getting my
Masters in Science and Engineering, (micro-electronics)
@ ESAT laboratory, KULeuven, Belgium.
----------------------------------------------------------------------
This archive was generated by hypermail 2b28 : Mon Jan 14 2002 - 13:30:15 EET