LC-221 date/time types

Dear Martin and Misha:

The W3C XML Schema Working Group has spent the last several months
working through the comments received from the public on the last-call
draft of the XML Schema specification.  We thank you for the comments
you made on our specification during our last-call comment period, and
want to make sure you know that all comments received during the
last-call comment period have been recorded in our last-call issues
list (http://www.w3.org/2000/05/12-xmlschema-lcissues).

Among other issues, you raised the point registered as issue LC-221,
which suggests various changes to the date/time types of XML Schema.

On some points, the WG agrees with you and has made changes where
appropriate.

   - The value space of the timePeriod is now described as a set of
     tuples, instead of a set of absolute time spans; the result is
     that the value space is closed under addition but not under
     subtraction.

   - We believe Appendix D is now correct.

   - Measurement units are not provided in XML Schema 1.0, but
     it is safe to believe that the issue of supporting them will
     be raised again in work on future versions.  We agree that
     it should be raised (not that that makes any difference: it
     will be raised whether you and we agree or not).

Some changes the WG declined to make:

   - Eliminating Gregorian months and years does not seem to us to
     increase the utility of XML Schema for anyone; retaining them does
     not seem to us to decrease the utility of XML Schema for anyone.
     On the contrary.  The current proposal seems to the WG to be
     somewhat closer to the 80/20 point than a version without months
     or years.  The Gregorian calendar is in fact in widespread use; we
     believe it is important to support it.  It would be useful to be
     able to support other calendars as well; we would like to see such
     support proposed as an alternative, perhaps for 1.1. We do not
     think that eliminating support for the dominant calendar is a good
     way of improving support for other calendars.

   - The restriction of timeDurations to units of days, hours, minutes,
     and seconds does not (as your comment appears to suggest) render
     timeDurations unambiguous.  The existence of leap seconds in the
     international standard calendar means that only timeDurations
     which completely lack seconds (as well as months and years) or are
     expressed only in seconds are wholly unambiguous as a multiple of
     their smallest unit without reference to a particular starting
     point.

   - Your general arguments against multiple lexical forms for simple
     types (discussed as issue LC-220) were not accepted by the WG.

   - That recurringDuration is of finite (i.e. limited) value we
     agree.  We do not believe that having finite value is the same as
     having no value; it seems to the WG to be useful enough to
     include.  We do not see any benefit from excluding it.  Some
     applications will require ways to express the third Wednesday
     of every month; they will require mechanisms outside of XML
     Schema 1.0 to do so.  Removing the current recurringDuration type
     would not affect this fact, except to require a larger group of
     applications to use mechanisms outside XML Schema 1.0.  We have
     been unable to discern any advantage in such a move.

   - You suggest that timeInstant should be "promoted to a base type".
     We believe you mean it should be made into a primitive type,
     instead of being a derived type.  We do not see why:  there is
     no honorific status in being primitive as opposed to derived.
     The type derivation relations among the builtin types are merely
     ways of expressing their subset and superset relations.

   - Ditto for your suggestion that date, too, should be made a
     base (primitive?) type.

   - The Gregorian calendar is a notation for writing dates.  The
     range of dates for which it provides a notation is not limited
     to the time periods during which it has served as the civil
     calendar in any particular jurisdiction.  Of course it's
     applicable to dates before 1582, in the sense that it has
     lexical forms denoting such dates.

It would be helpful to us to know whether you are satisfied with the
decision taken by the WG on this issue, or wish your dissent from the
WG's decision to be recorded for consideration by the Director of
the W3C.

with best regards,

-C. M. Sperberg-McQueen
  World Wide Web Consortium
  Co-chair, W3C XML Schema WG

Received on Thursday, 5 October 2000 21:50:20 UTC