Messages in this thread |  | | From | Alan Cox <> | Subject | Re: Kernels - What's in and what's out. | Date | Thu, 9 May 1996 09:55:03 +0100 (BST) |
| |
> You need some sort of hook in the kernel. The bulk of the voice > recognition software can go in user space, but the kernel needs > to supply some way to put characters into the keyboard buffer.
What do you think pty's are for ;) - how do you think screen and splitvt work.
> The bulk of the code should be a real-time process locked into > memory. You wouldn't want your voice software to get swapped out.
Oh I would. So long as the data is buffered I'll let it swap back in and think thank you.
|  |