Subject: Re: New read & write syscalls
To: Bill Studenmund <email@example.com>
From: Bill Sommerfeld <firstname.lastname@example.org>
Date: 06/30/1999 17:22:18
At a purely conceptual level, I think the idea of an "alternate open"
rather than an "alternate read/write" is likely to be much cleaner.
For one, some implementations might conceivably want to use stdio to
write to the underlying layer..
There are more worked examples in the system of "alternate open"
(e.g., socket(), accept()) than "alternate write" (about the only case
I can think of is out-of-band data).