Lawrence D'Oliveiro
2024-05-22 02:24:38 UTC
One of the recent additions to the Bitsavers vintage computing
collection is this presentation from 1986
<http://bitsavers.trailing-edge.com/pdf/ibm/share/SHARE67_Presentation_on_Setting_a_Computers_Clock_Correctly_1986.pdf>
at a session of SHARE (one of the IBM user groups), on how to ensure
your computer’s time is set correctly.
Remember, this was before NTP. But there was already a worldwide
network of atomic clocks, keeping synchronization with each other, and
there were shortwave stations, like WWV in Boulder, Colorado, that
would broadcast pips once per second, synchronized to the atomic clock
network. And there were even consumer-level gadgets, like the one
mentioned in the article from Heath, that would automatically decode
those broadcasts and output a time code a computer could read.
The article has some nice background info. Like did you know there can
be up to a 1-second discrepancy between GMT and UTC? (Actually I think
from other sources the limit might be more like 0.9 seconds.)
IBM’s mainframe OS developers were perceptive enough to realize that
it can be useful to have the computer keep time in GMT, and maintain
an offset for calculating local time. Particular site admins could of
course have the choice of using local time for the system time, and
setting the offset to zero.
Changing the offset for daylight saving could be a fraught procedure,
though. The level of confusion it could cause for running applications
was such that it was safer to do a reboot.
One little item I don’t recall coming across before, another of those
NASA programming screwups:
One of the famous programming mistakes of all time was by the
programmer who thought the earth revolved on its axis once every
24 hours and used that fact to calculate landing sites for the
Mercury space flights. Our first astronauts spent a lot more time
in the water than they should have because of that.
collection is this presentation from 1986
<http://bitsavers.trailing-edge.com/pdf/ibm/share/SHARE67_Presentation_on_Setting_a_Computers_Clock_Correctly_1986.pdf>
at a session of SHARE (one of the IBM user groups), on how to ensure
your computer’s time is set correctly.
Remember, this was before NTP. But there was already a worldwide
network of atomic clocks, keeping synchronization with each other, and
there were shortwave stations, like WWV in Boulder, Colorado, that
would broadcast pips once per second, synchronized to the atomic clock
network. And there were even consumer-level gadgets, like the one
mentioned in the article from Heath, that would automatically decode
those broadcasts and output a time code a computer could read.
The article has some nice background info. Like did you know there can
be up to a 1-second discrepancy between GMT and UTC? (Actually I think
from other sources the limit might be more like 0.9 seconds.)
IBM’s mainframe OS developers were perceptive enough to realize that
it can be useful to have the computer keep time in GMT, and maintain
an offset for calculating local time. Particular site admins could of
course have the choice of using local time for the system time, and
setting the offset to zero.
Changing the offset for daylight saving could be a fraught procedure,
though. The level of confusion it could cause for running applications
was such that it was safer to do a reboot.
One little item I don’t recall coming across before, another of those
NASA programming screwups:
One of the famous programming mistakes of all time was by the
programmer who thought the earth revolved on its axis once every
24 hours and used that fact to calculate landing sites for the
Mercury space flights. Our first astronauts spent a lot more time
in the water than they should have because of that.