Native code implementation.

[Viktor L]

I'm currently playing around with sound in D, using the native Linux
functionality (eg. using ioctls, write and read and so forth) to do so.

I'm wondering if I'm doing it properly, design wise. I've been doing it sort
of like I would if i were to write a C program that uses lots of assembly,
confining it to wrapper functions (well, wrapper classes in this case) so that
the rest of the code is as clean from native functions as possible (this way I
imagine porting to other systems is easier).

Is this the best way of doing this, or should I consider other ideas?

Thanks.