martin.burnicki at meinberg.de
Mon Jul 7 12:57:14 UTC 2008
David Woolley wrote:
> You also need to be in a country that sets rules algorithmically.
> Whilst the UK now uses a week 5 rule, it wasn't that long ago that the
> rules were set by secondary legislation each year, and generally
> differed from those in the primary legislation. More recently, Pakistan
> introduced DST with only two weeks notice, and appear to have only done
> it for one year. Even the best maintained site would have said there
> was no DST if you ran the algoritthm before the middle of May, and most
> probably wouldn't give the right result until some time in the middle of
> June. The resulting rule would not be valid from June next year, unless
> new legislation is introduced. The rule is probably an Mn.0.1 rule, but
> one can only guess that from a sample of one.
As already stated earlier IMO we have to distinguish between the
possibilities a set of rules provides, and the policies determined by the
If a country only publishes a rule 2 weeks before it comes into effect then
you have to update both you timezone libs or your rule string quickly, or
the DST changeover will not happen correctly. Similar for a rule which can
not easily be covered by those standard rules (first Sunday after ..), e.g.
in case of Israel.
In this context there is no difference whether you use libs or a rule
More information about the questions