Re: ada calendar

[Nick Roberts <nick.roberts@acm.org>]

tmoran@acm.org wrote:

>>Will this revision include definitions based on precise
>>description of movements of the Earth ?  The rotation
> 
>    My understanding is that Ada.Calendar is basically a wall clock - it
> can be reset by the operator for Daylight Savings, leap seconds, clock
> skew, or whatever.  It's the right thing to use for printing a date & time
> on a report, for instance.  Ada.Real_Time has a monotonic clock so it's
> generally the right thing for elapsed time.  Inserting a leap second for
> instance would cause the clock to run backwards one second, which would be
> illegal for Ada.Real_Time.Clock, but perfectly OK for Ada.Calendar.Clock

I think this is quite correct. The proposal does include minimal (very 
minimal) time zone functionality, but I suspect this might be removed by 
the ARG (or WG9), as being too horrible to contemplate.

[For one thing, it is only the rather obscure UT0 time system that is 
actually based on movements of the Earth; UT1 corrects it for 'wobble'. UTC 
is based on atomic clocks, and is corrected in steps of precisely 1 second 
every now and then to bring it within 0.9 s of UT1. As for the all the 
different time zones in use around the world, that is a deep mire where 
only the very foolish or the very brave (or those who enjoy maintaining 
huge databases) dare to tread.]

-- 
Nick Roberts