[ih] Origin of 'talk' command

Jerome H Saltzer Saltzer at MIT.EDU
Mon Dec 23 11:05:58 PST 2002

At 1:15 PM -0500 12/23/02, Bob Frankston wrote:
>Thanks. I don't know the details of CTSS but judging from Multics and
>CP/CMS, yes, it was very different. There seemed to be two approaches.
>The Dartmouth/CTSS "outsourced" TTY handling and the main system
>processed lines of text whereas the 940 and DEC systems (derived, I
>think from the 940 thinking) had no qualms about processing each
>character. Over time, of course, the approach converged.

CTSS had options of character-at-a-time or line-at-a-time processing, but
waking up the user VM on each character was so expensive that the feature
never had much use except in experiments.

>From the CTSS programmer's guide description, I think that one fundamental
Status: RO

difference between it and the 940 system is that in CTSS inter-user talk
was done by copying the sender's message into the recipient's *input*
buffer, where whatever program was listening for input would see it (and
potentially massage it) before echoing it to the recipient's terminal.  If
Butler's recall of the 940 system is correct, the sender's output message
was copied into the recipient's *output* buffer, so it went directly to the
terminal and the recipient's input listener didn't get a copy.


More information about the Internet-history mailing list