Re: CLDR proposal: Holiday rules

From: Stephen Colebourne (
Date: Thu Oct 20 2005 - 17:15:12 CST

  • Next message: David Starner: "Re: LAO LETTER FO SUNG... wikipedia"

    Philippe Verdy wrote:
    > I would also like to have such data, but it seems difficult to be
    > exhaustive here, notably because these dates are often complex to
    > compute, and tend to use various calendar rules, and have lots of
    > exceptions. To be complete, the data would need to include history, and
    > as well support a set of external symbolic rules (for example the
    > various ways to compute the Easter day, some cosmologic rules based on
    > actual or historical rules, and some other religious rules, some of them
    > being impossible to compute precisely as the effective dates are
    > determined each year by some authority, for example the exact date for
    > the start of Ramadan, which not only depends on calendar, but also
    > depends on location on earth!).

    Our initial start already includes a relative concept. That is to be
    able to define certain rules using code (eg Gregorian Easter), and then
    reference other rules relative to that using XML (eg Whit Monday, 50
    days later).

    I currently believe that with the right base definitions, such as
    GregorianEaster, or OrthodoxEaster (coded, not defined in XML) the rest
    is achievable. For non Gregorian calendars etc, defining those base
    definitions will be key.

    > Finally, such data would need to specify the status of such dates: for
    > example, a data may be a holiday only for some category of population,
    > but not all. In France we have the concept of legal holidays that must
    > be worked, unless there's a special organization-level agreement that
    > moves that worked holiday to another holiday, except for some legal
    > holidays that apply for everybody (such as the New Years Day and the
    > National Day that US designates as Bastille Day).
    > The status of all these days can be complex to determine exhaustively:
    > whever a day is worked or not depends on national legislation, and
    > organizations. Additionally, there may exist "extra" days that don not
    > exist every year (for example if some normally worked monday falls
    > between a normally non-worked sunday and a non-worked holiday) that
    > cause some businesses to be closed as well, or have limited activity
    > (hink about stocks exchanges). Whever these days will exist may vary
    > over time, and often without legal status, depending on rules applied in
    > each organization.

    Indeed, so part of the aim of this is to define an XML structure that
    companies can define their own rules into. This is definitely not just
    locale specific data.


    This archive was generated by hypermail 2.1.5 : Thu Oct 20 2005 - 17:16:08 CST