[tz] Vadsø, Norway mayor wants 26-hour days

[Paul Eggert]

On March 20 Wenche Pedersen, the mayor of Vadsø municipality, Norway, 
asked[1] the European Commission to issue a directive allowing Norway to 
create a separate time zone for the northern part of Norway. Days in the 
new time zone would contain 26 hours instead of the usual 24.

This request is part of the "MOREtime" project. Mayor Pedersen wrote:

> ... we aim to celebrate and promote this unique way of life, offering individuals the opportunity to enjoy more quality time engaging in activities such as fishing, hunting, learning new languages, or simply being with loved ones. By having our own certified time zone within the EU, we hope to attract more young people, not only from Norway but also from around the world, to discover and settle in this adventurous area. 

Recently the mayor's proposal has been covered by several major news 
organizations. When asked how the 26-hour day would actually work, the 
mayor replied "We haven't thought a lot about that. The clock will go 
from 12 to 13 ... and we have to see how this will go. I don't think 
they're going to say yes so we haven't thought about all the details."[2]

Longtime readers of the time zone mailing list might remember that in 
2019 a resident of Sommarøy, Norway petitioned his parliamentarian to 
abolish civil time in his town, something that also made the news.[3] 
However, the Sommarøy proposal was by a private individual and was 
easily implementable via tzdata by creating a new UTC or LMT zone with 
abbreviation "-00" or "LMT", whereas the Vadsø proposal is by the mayor, 
and would require changes to tzcode and would therefore be considerably 
harder to support.

Although it's unlikely that the Vadsø proposal will go anywhere, it's 
helpful to think of how timekeeping changes like this might be 
accommodated by the TZDB project, as similar proposals might happen in 
the future. I see several ways to implement 26-hour days, falling into 
two major categories.


Category I: Shorten civil-time hours so that there are 26 civil-time 
hours in a standard-length (SI) day. Four possible ways to do this are:

A. Civil-time seconds are 12/13 s (i.e., 12/13 of an SI second, or about 
0.923 s), so that civil-time minutes and hours are both 12/13 of their 
SI counterparts.

B. Civil-time seconds are the same as before, but civil-time minutes are 
60 * 12/13 s (i.e., about 55.38 s).

C. Civil-time seconds and minutes are the same as before, but civil-time 
hours are 60 * 12/13 min (i.e., about 55.38 SI minutes).

D. Civil-time minutes are 60 * sqrt(12/13) s (about 57.65 s), and 
civil-time hours are 60 * sqrt(12/13) civil-time minutes (i.e., about 
57.65 civil-time minutes).


Category II: Adjust day lengths so that all or most days contain 26 SI 
hours. Three possibilities here are:

E. Civil-time days are 26 SI hours, and the calendar gradually falls 
further and further behind the Gregorian calendar.

F. Civil-time days are 26 SI hours, but skip every 13th day so that days 
stay roughly synchronized with the Gregorian calendar.

G. At the end of every day, move clocks backward two hours, from 24:00 
to 22:00, so that timestamps from 22:00:00 through 23:59:59 repeat. Do 
this consistently for eleven days, thus moving clocks backward a total 
of 22 hours. On the twelfth day move clocks forward 22 hours instead, 
from 24:00 to 22:00 the next day. Repeat this cycle indefinitely. This 
way, eleven days out of thirteen contain 26 hours, and the next two days 
combined also contain 26 hours - 24 hours in the twelfth day and 2 hours 
in the thirteenth.


Of these methods something like (G) would be the most compatible with 
computerized timekeeping, as (G) conforms to POSIX-1.202x Draft 4.1[4] 
whereas (A) through (F) do not. (G) avoids timestamps like 25:30; G's 
hours, minutes or seconds are all of the usual lengths; and (G) contains 
only transitions that are like daylight-saving transitions that 
computerized timekeeping can already deal with.

Although a downside of (G) is that at times the sun will be at its 
zenith at midnight, I assume Vadsø residents are already accustomed to 
midnight suns so this disadvantage is relatively minor for them. Another 
downside of (G) is that clocks will never go "from 12 to 13" as the 
mayor requested, but given the major technical difficulties of (A) 
through (F) I hope the mayor would appreciate that (G) may be the best 
we can practically hope for in today's standardized and computerized world.

Unfortunately even (G) would require some changes to tzcode and tzdata 
to become practical, and it would take many years for these changes to 
propagate to devices in common use in Vadsø. So, unless Vadsø residents 
are prepared to change their clocks by hand every day for the next 
several years, it does not appear that the mayor's proposal will be 
practical any time soon.

I am sending this email to the time zone mailing list, as perhaps 
someone on the list can think of even better ways to implement the 
mayor's request. I am also cc'ing this email to Mayor Pedersen to give 
her a heads-up about this analysis, should it ever be needed in Vadsø.

[1]: https://kommunikasjon.ntb.no/files/17848798/18066114/57109/no
[2]: 
https://www.politico.eu/article/norway-arctic-region-asks-eu-commission-for-26-hour-day/
[3]: https://mm.icann.org/pipermail/tz/2019-June/028154.html
[4]: https://www.opengroup.org/austin/