Re: [LEAPSECS] Consensus rather than compromise

From: John.Cowan <jcowan_at_REUTERSHEALTH.COM>
Date: Wed, 31 Aug 2005 22:57:42 -0400

Rob Seaman scripsit:

> 2) It is different because today timezones and related issues like
> Daylight Saving are purely periodic phenomena. For most purposes -
> including interval purposes - you only need to know the zone and DST
> status at the beginning and end of whatever period of interest.

That turns out not to be the case.

> 3) It is different because today there is a single list of historical
> leap seconds, and a TZ based system, as Cowan points out, would be at
> the whims of local jurisdictions worldwide.

It already is.

> One imagines a new discipline similar to the tree ring counting of
> dendrochronology whose goal would be to synchronize timestamped data
> from different localities and epochs. These chronochronologists would
> be responsible for building overlapping longitudinal historical clock
> records from hundreds of distinct locations worldwide.

Fortunately, it's not that bad. All 365 current zones define
their LCTs using a list of offsets from a common reference time.
For each such offset, we must keep track of whether it is a standard or
daylight-savings offset and during what range of moments it was in effect.
That's essentially what the files under /usr/share/zoneinfo contain,
in binary form.

Consequently, given a particular moment in one time zone given as
year-month-day-hour-minute-second plus an indication of whether the
moment is in DST or not (to compensate for the fact that some LCT
clock readings labels refer to two separate moments), you can generate
equivalent information for every other zone without problem.

> 4) It is different precisely because for short periods the existence
> of leap seconds can often be ignored for purposes that require
> accuracy at the level of a minute or so. Presumably a TZ/DST based
> system would operate in larger discontinuities, e.g., leap hours,
> that might be very difficult to ignore indeed.

Agreed, which is why people particularly concerned with a given zone
would be at pains to update their information in a timely fashion.
Systems that don't care what LCT zone they are in wouldn't have
to worry about it.

John Cowan
This great college [Trinity], of this ancient university [Cambridge],
has seen some strange sights. It has seen Wordsworth drunk and Porson
sober. And here am I, a better poet than Porson, and a better scholar
than Wordsworth, somewhere betwixt and between.  --A.E. Housman
Received on Wed Aug 31 2005 - 19:59:42 PDT

This archive was generated by hypermail 2.3.0 : Sat Sep 04 2010 - 09:44:54 PDT