overlapping of rules? - Palestine

[Michael Deckers]

    On 2010-10-13 18:29, Arthur David Olson wrote
    on superfluous time zone rules:

 >  (Note that the time zone compiler notices that nothing
 >  changes on the first Friday of September, 2010 and does
 >  not produce any "extra" output.)

    Yes, obviously a switch to winter time directly after
    another switch to winter time should not hurt. But
    such  "overlapping" switches may confuse a human
    reader, and in some (imaginary) cases, even the
    compiler.

    Consider a rule like
       # Rule  NAME   FROM TO  TYPE IN  ON      AT   SAVE
       Rule Palestine 2009 max -    Mar lastFri 0:00 1
    It says that summer time is used from its value
    [a]    2009-03-27 + 01 h onwards
    if winter time has applied before that instant. If
    summer time had already applied before that instant,
    then the rule would say that summer time is used
    from its value
    [b]    2009-03-27 + 00 h onwards
    which is, in fact, one hour earlier.

    Together with the (fictitious) "overlapping" rule
       # Rule  NAME   FROM TO   TYPE IN  ON  AT    SAVE
       Rule Palestine 2009 only -    Mar 26  23:30 1
    which specifies summer time from
    [c]    2009-03-27 + 00:30 h onwards,
    the interpretation becomes doubtful: if the
    second rule [c] is honored, then [a] cannot really
    apply because local time never took the value
    2009-03-27 + 00 h; but [b] implies that [c]
    cannot apply because local time could never
    have taken the value 2009-03-26 + 23:30 h.

    I am not saying that this is a real problem --
    I am just proposing to avoid any "overlapping"
    (useless and potentially confusing) rules.

    Michael Deckers.