David Holland <dholland@eecs.harvard.edu> writes:
> When was the equivalent test put into write? Ages ago.

Yup.  what do commercial unixes do with write?


> Is there any
> reason (besides inertia) that wall should be different?

(1) inertia (a.k.a. consistency with other UNIXes) is not to be
discounted.  In the absence of an actual real reason to change, it
_should_ be valued over arbitrary change.  (If it ain't broken...)

(2) it's not obvious to me that your solution has much real use given
the way (at least I think) these tools are used.

At least the way I think about it, 'write' is used to communicate with
specific people, i.e. a conversation (perhaps via write, or talk, or a
phone call) is expected to ensue.  In other words, if somebody can't
reply, that's a problem (hence the check).

Wall, on the other hand, is used to communicate status to lots of
people, typically where conversation is _not_ expected to follow, as
far as i'm concerned.  (yes, there are certainly exceptions to this,
but it's probably the norm.)  If you're not expecting a reply, why
bother warning if you can't get one?



cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/Pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.