Eugene van der Pijll wrote: > Lord Tanlaw refers to the UK as the only >industrialized country that has its legal time based on GMT. But there >are probably a lot of former colonies that have not changed their time >laws.
Found a source for more on this. <http://www.ucolick.org/~sla/leapsecs/ onlinebib.html> has a section about it, with many links. It explicitly lists some countries where legal time is based on UTC: France, Germany, Hong Kong, Korea, Netherlands, New Zealand, Sweden, and Switzerland. It also explicitly lists some that are based on GMT: the European Union, Ireland, Namibia, United Kingdom, United States (former colony!), and the canadian provinces Ontario, Saskatchewan, and Quebec. So the noble Lord was in error on that point. The context for that bibliography page is the debate over whether leap seconds should continue. One of the problems that would occur if UTC were to have no more leaps is that UTC+offset would cease to be a good approximation of legal time in countries where legal time is GMT+offset. It's essentially impossible to get all the countries in the world to legislate on anything. I also searched the Olson database. As we know it doesn't directly address this topic, but I looked for offsets that were not in integral minutes. A fractional-minute offset is a dead giveaway for use of solar time, because UTC can't accept such offsets. The database doesn't list any timezone with a current offset involving fractional minutes. The last usage of such an offset that it records was in Liberia, and ceased on 1972-05-01. The database doesn't actually say much about Saudi Arabia; the apparent-solar-time files are presented without real explanation. I'd like to introduce for consideration the concept of explicitly vague Universal Time. The term "UT" on its own is ambiguous: it refers to UT1, UT2, UTC, UT0, UTC-SLS, and others, collectively. If one is working at the minute level or above, the differences between the flavours of UT are insignificant, so merely specifying "UT" suffices. Unix timestamps that predate precise UTC synchronisation are ambiguous in just this way, and so Unix time in that era is best understood as being based on "UT", rather than UTC or GMT specifically. This seems to be the best way to think about DateTime's current time scale too, when working outside the range of known leap seconds. I think this vague UT is what we should consider Olson timezones to be based on, not UTC. I asked the tz mailing list about base time scales a while ago, and their answer supported this interpretation. In the present era they consider the timezones to be based on "UTC" (that's the term they use), but they don't really think about the meaning of that, and they have no answer for what time scale they're based on prior to UTC's existence. It's the usual problem with interdisciplinary timekeeping: three or four orders of magnitude difference is enough to prevent most meaningful dialogue. With that interpretation of the Olson database, an implementation is free to use the database's offsets with whatever form of UT it finds most convenient: most likely UTC. Until the database does properly address the issue, let's not perceive information that's not there. -zefram
