On Wed, 17 Jun 2020 07:33:48 -0700, Charles Mills wrote:
>If a mainframe process is synchronizing with an off-box process, or engaging
>in communications that contains a timestamp governed by a standard, isn't
>"true" UTC a requirement?
>
The current CVTLSO is 27 seconds:
https://datacenter.iers.org/data/latestVersion/16_BULLETIN_C16.txt
(See Principles of Operation for explanation.)
In 27 seconds an airliner travels over 6 km. That can matter.
>-----Original Message-----
>From: Peter Relson
>Sent: Wednesday, June 17, 2020 5:38 AM
>
>Is it worth reminding that leap second offset and time zone are basically
>for human consumption?
>
In the end, all products of IT are for human consumption. The customers
pay the bills.
>The values used for real processing ought to be the value returned by
>STCK/STCKE/STCKF unmodified.
>
I strongly agree. But where is the boundary between "human consumption"
and "real processing"? For example, why are ISPF timestamps in local time?
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
