On 2020-06-10 21:25, Paul Eggert wrote:
Anyway, the bottom line is that the stopped clocks make this an oddball transition that cannot be modeled exactly by tzdb, and that given the story you mentioned you are correct that it's better modeled using an ordinary transition (from 00:09:21 to 00:00:00) than the unusual transition we're currently using.
Let us not get carried away and consider this an oddball transition. It was a transition from a local time to a time zone time that is just a few minutes slower, as has happened in many locations all over the world. Legal time in France was never stopped; it just switched from UT + 00:09:21 to UT. Since 1891-03-16 (sic!), legal time in France was the mean solar time in Paris, and the law published on 1911-03-10 (online at [https://gallica.bnf.fr/ark:/12148/bpt6k2022333z/f2]) changed it to mean solar time in Paris minus 00:09:21; this change took effect on the day after publication, 1911-03-11. The appended decree (loc cit) makes it clear that the new definition of legal time was used with values from 1911-03-11T00:00:00 onwards. There is no mention in the legal texts of how the change is to be effected (let alone that legal time would stop for some time). Hence, the old definition of legal time applied until just before it took the value 1911-03-11T00:09:21, and the jump in legal time was from 1911-03-11T00:09:21 to 1911-03-11T00:00:00. The newspaper reports about some remote-controlled public clocks ("pendules pneumatiques") shows that these clocks were stopped when they reached 1911-03-11T00:00; the next minute pulse only came 10 minutes and 21 seconds later and put them to 1911-03-11T00:01. That does not imply that legal time had stopped, it just indicates how these clocks were adjusted. Other clocks (such as clocks on churches and pocket watches) would certainly have been adjusted in a less time-consuming manner. Railway time in France was also adjusted at the same occasion from UT + 00:04:21 to UT. Here it may be more appropriate to say that "time was stopped" (at least for the trains not currently in motion) because railway time is the parameter used for planning train motions and should be monotone. Michael Deckers.