> I've looked at it, and here's what I can tell you:
> 
> * Porting this software to NetBSD would be a lot of work. It would
>   require a significant knowledge of the internals of character device
>   drivers. If you were to successfully complete this port, you would
>   be considered a successful kernel hacker. It is a somewhat beyond my
>   ability at least.

I might just try and learn whatever I need on the way !

> * There is also a legal issue. The README file states that all the 
>   code is either licensed under the GPL or "terms similar to the
>   BSD license." This is flat out wrong. A few (Linux-derived) files 
>   are licensed under the GPL. For the rest, the LICENSE file states: 
> 
>   [...]  Permitted use. Redistribution and use in source and binary
>   forms, without modification, are permitted under the terms set forth
>   herein.

>   This means you could not modify any of their code. You would have to
>   start from scratch. As a matter of law, [Note: IANAL] I belive it
>   would be legal to use _facts_ that you deduced from reading their
>   software to write your own; nevertheless, they could make your life
>   very hard(read: sue you and make you go to court) if they wanted to.

That looks similar to the early time of BSD 4.4 and BSDLite 4.4. Anyhow,
the driver code won't be the same, obviously ! I don't figure out how
it would not be rewritten from scratch, since Linux and BSD are quite
different at that level. As for the DSP objects files, they don't need
any modification, so they comply with the terms of the licence.

By the way, if I faintly remember, any software using even parts of GPL
code must be wholly GPL'ed itself, no ?

That is a mistery to me : you buy a device, hardware, though thing, and
you get with it a software that obviously cannot work without the
hardware. Admitting that making a hardware copy of a card is thousand
times more difficult than copying lines of code, what is the point in
licencing the software ? And then, why hinder porting, since it
potentially broaden the number of people you can sell you device to ?

> You might try contacting them and asking them if they would support a
> porting effort to NetBSD. You might also ask them if anyone from 
> another BSD project has contacted them. 

I've sent an email. Never got any answer. I suspect the guy is alone. I
send out an email in French, given that the guy seems to be a French
canadien, but I now begin to doubt. :)

> In short, the situation is not too good. Writing device drivers that 
> use proprietary object code that will only ever run on one platform is
> not a high priority for NetBSD kernel people. (If you are willing to
> pay for this, you might get something done. You would be looking at a
> a bounty of at least $500 or equivalent). 

Paying for something you can do yourself is a waste of material :) I
consider that learning how the kernel works is largely worth my time and
money ! And I don't speak about jet lag :)

Vincent