Origin Of The Autobaud Technique

[Lawrence D'Oliveiro]

I can remember when our DEC VAX/VMS system was upgraded sometime in the
early 1980s to version 4, I think it was, which introduced a clever new
terminal-line feature called “autobaud”. Gone was the pain of sitting down
at a terminal and watch it spew rubbish until you figured out what speed
the last user had left it at: of course they would have had both terminal
and line at the same speed, but if this was a VT100-class terminal, and
they had neglected to save the terminal speed change to its nonvolatile
RAM, then if the terminal was turned off and on again, it would revert to
its previous setting, which would be out of sync with the line.

Instead, the docs said you just had to press RETURN once or twice, and the
terminal driver would automatically detect the right line speed and pop up
a nice, legible login prompt. In practice, I don’t recall ever having to
press RETURN more than once. It just seemed like magic. (We had fewer TV
channels in those days...)

Was DEC the first with this? Somehow I suspect not ...

[Lynn Wheeler]

Post by Lawrence D'Oliveiro
Instead, the docs said you just had to press RETURN once or twice, and the
terminal driver would automatically detect the right line speed and pop up
a nice, legible login prompt. In practice, I don’t recall ever having to
press RETURN more than once. It just seemed like magic. (We had fewer TV
channels in those days...)

... mentioned it in post to
Any interesting PDP/TECO photos out there?

thread a few hrs ago ... in the late 60s, did it for clone ibm 360
telecommunication controller we built using Interdata/3 machine
(upgraded to Interdata/4 with cluster of Interdata/3s) ... that
Interdata and then Perkin/Elmer sold.
https://en.wikipedia.org/wiki/Interdata
https://en.wikipedia.org/wiki/Perkin-Elmer#Computer_Systems_Division

initial (virtual machine) CP/67 delivered to univ had 1050&2741 terminal
type support with automagic terminal type recognition. Univ. had ascii
TTY (mostly 33, but some 35), so I added ascii terminal support
integrated with automagic terminal type recognition (able to use the SAD
CCW to switch the terminal type line scanner for each line/port). I then
wanted a single dial-in number ("hunt group") for all terminal types
... but while the terminal type line scanner could be switched for each
port, IBM had hard-wired the port line speed ... thus kicked-off the
univ project to build our own clone controller that also did "autobaud".

[Lynn Wheeler]

Post by Lynn Wheeler
initial (virtual machine) CP/67 delivered to univ had 1050&2741
terminal type support with automagic terminal type
recognition. Univ. had ascii TTY (mostly 33, but some 35), so I added
ascii terminal support integrated with automagic terminal type
recognition (able to use the SAD CCW to switch the terminal type line
scanner for each line/port). I then wanted a single dial-in number
("hunt group") for all terminal types ... but while the terminal type
line scanner could be switched for each port, IBM had hard-wired the
port line speed ... thus kicked-off the univ project to build our own
clone controller that also did "autobaud".

trivia: turn of century had tour of datacenter that handled majority of
dial-up POS credit card swipe terminal calls east of the mississippi ...
the telecommunication controller was descendant of what we had done in
the 60s ... some question that the mainframe channel interface card was
same design we had done more than three decades earlier.