Which tzdata zones have complete pre-1970 history?

[Alois Treindl]

I think it would be a good start investigating for which of the approx.
400 tzdata zones the pre-1970 history can be considered complete and
correct.

We have no absolute reliable measure what 'complete and correct' means.
But a comparison with a recent revision of Shanks/Pottenger American and
International atlas would already serve to a large extent.

As a working definition I propose:

Those tzdata zones whose history details are equivalent or better than
the corresponding history lists by Shanks & Pottenger should be
considered complete and correct.

Example:
Switzerland
Switzerland is not historically complete in tzdata.

tzdata zone Europe/Zurich
is correct, but a separate zone Europe/Geneva is missing.
The canton of Geneva had Geneva mean time 6E0915 between 1851 and the
introduction of standard time on 1 June 1894
(SOURCE: GABRIEL, TRAITE DE L'HEURE DANS LE MONDE)

Shanks/Pottenger have a separate table for this canton, but it contains
a small error as well: it assigns LMT until 1894, while in fact it
should be Geneva city observatory mean time (24m37s east of Greenwich)
between 1851 and 1894. The error is small because this canton is not
big, and the most distance any other town has from the Geneva meridian
is 4 minutes of arc.

[aloist]

My studies showed that not only the cities, but the whole areas defined in TZ documentatin as belonging to the zone
are correctly represented in their time zone history since the beginnning of standard time.

The analysis is based by comparing TZ data with all other reference works know to me, and used in astro.com services.

These are 227 out of the 384 zones as defined by tzdata2019b

The list zone.txt is attached

[David Patte]

I am a member of the tz mailing list itself, and this topic has been often discussed.

The pre-1970 tz data is only maintained for the mentioned city for each current zone.

For example, Montreal had some different rules before 1970 when compared to Toronto, though both cities use the Toronto tz data at the current time, since they have had the same rules since 1970.

As far as I know, there is no general effort to maintain pre-1970 rules for other cities.

--
--
You received this message because you are subscribed to the Google
Groups "tzdata history" group.
To post to this group, send email to tzdata-...@googlegroups.com
To unsubscribe from this group, send email to
tzdata-histor...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/tzdata-history

---
You received this message because you are subscribed to the Google Groups "tzdata history" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tzdata-histor...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tzdata-history/746da8c3-dd52-411a-aa50-3aceadd79de6%40googlegroups.com.

[Alois Treindl]

I know.

I wanted to say, but did maybe not say it clearly enough:

For those zones in my list, 225 out of approx. 400, the TZ rules represent the whole area before 1970.

For the other, approx. 175 zones in TZ, which I did not list, the TZ zone rules do not represent the whole area before 1970, only the area past 1970.

About 2000 extra zones are needed to represent the details of the pre-1970 history. As far as kniwn to me.

I know that TZ will never aim to cover this. It is up to enthousiasts outside of TZ to do it privately. 

The main problem is the description of the area boundaries. TZ avoids that by limiting itself to reference cities and being vague in the area definition. Like what belongs to the rules of Europe/Paris when? All of France after 1970, sure, but what about 1940 to 1944 or 1914 to 1918?

The cause of the complexity are mostly frontier and front movements in both world wars in Europe, the wild DST irregularities in US and Canada before 1967 and complex local history in Indonesia. 

Africa and South America are well covered by the TZ zones for nearly all time.

To view this discussion on the web visit https://groups.google.com/d/msgid/tzdata-history/7eb267aa-2d16-f4f5-4bdb-62c68def333c%40relativedata.com.

[David Patte]

Thanks for the clarification.

It might be worthwhile to make a list of those zones where we in fact
know for sure that they do not represent the whole current area, such as
Toronto.

Like you, I am also interested in pre-1970 timezone rules. I have
written a parser of the raw tz data for past-date astronomical
calculations I use in my own astronomy apps.

Since I have your attention, do you use a general cutoff date for when
people switched from apparent solar time to mean solar time to record
events?