Loading ...
Sorry, an error occurred while loading the content.
 

Re: [ISO8601] Are time intervals inclusive/exclusive?

Expand Messages
  • Tex Texin
    Paul, Thanks for the reply. 1) Using 3.17, I would presume the endpoints are not included (because it says between ). However, by 3.26 the endpoints of the
    Message 1 of 4 , Sep 19, 2001
      Paul,

      Thanks for the reply.

      1) Using 3.17, I would presume the endpoints are not included (because
      it says "between").
      However, by 3.26 the endpoints of the interval are separated from the
      two time-points by the width of a dimensionless instant, so practically,
      I can include them... (Meaning, I physically start and end the clock
      concurrent with the time-points, and the next interval starts at the
      same time the previous one ends.)

      2) Although a duration is defined by time-points the spec also allows
      for reduced precision, so I don't see that using the date format alone
      necessarily implies the start of the day. I take it to mean the
      precision is reduced and you don't know where in the day the duration
      begins or ends.

      This is an important distinction. If using reduced precision implied a
      more granular time point, then it could cause problems for applications
      that depended on the reduced precision of the value. For example, if I
      tell you I will send a message on 2001-09-20, and you assume that since
      no message arrived by midnight 2001-09-20, it is ok to cancel the
      transaction, and the message shows up some time during 2001-09-20, we
      would have a dispute...

      3) I assume laps means lapse. I see it is used in multiple places in the
      spec, but not in my dictionary.
      ;-)

      tex


      P A Hill & E V Goodall wrote:
      >
      > Tex Texin wrote:
      > >
      > > 1) I don't see in the specification whether intervals include or exclude
      > > the endpoints from an interval.
      >
      > "3.17 period of time (time-interval)
      > portion of time between two time points"
      >
      > "3.26 time-point
      > instant in the laps of time regarded as dimensionless"
      >
      > Reading through the standard I see that this technical time-point term
      > is not used extensively in the document.
      >
      > This means that when you say 2001-09-11 in a duration expression you
      > are not referring to the whole day, but the start of the day of 2001-09-11.
      >
      > The examples shown for time-intervals identified by start and end support
      > this idea of time-point even if the term is not used.
      >
      > From
      > "5.5.4.1 Representation of time-intervals identified by start and end"
      > ...
      > "Extended format: YYYY-MM-DDThh:mm:ss/YYYY-MM-DDThh:mm:ss
      > EXAMPLE 1985-04-12T23:20:50/1985-06-25T10:30:00
      > A time-interval
      > beginning at 20 minutes and 50 seconds past 23 hours on 12 April 1985 local time
      > and ending at 30 minutes past 10 hours on 25 June 1985 local time."
      >
      > Since recurring time-intervals can use the above as part of specifying a
      > recurring time-interval the translation of the interval into words indicating
      > the use of time-points is directly applicable.
      >
      > > 2) As long as I am at it let me ask if anyone else considers the word
      > > recurrence ambiguous.
      >
      > I agree with you, I would have used the word "occurrence".
      >
      > Main Entry: oc·cur·rence
      > Pronunciation: &-'k&r-&n(t)s, -'k&-r&n(t)s
      > Function: noun
      > Date: 1539
      > 1 : something that occurs <a startling occurrence>
      > ----> 2 : the action or instance of occurring
      >
      > Main Entry: re·cur
      > Pronunciation: ri-'k&r
      > Function: intransitive verb
      > Inflected Form(s): re·curred; re·cur·ring
      > Etymology: Middle English recurren to return, from Latin recurrere,
      > literally, to run back, from re- + currere to run -- more at CAR
      > Date: 1529
      > 1 : to have recourse : RESORT
      > 2 : to go back in thought or discourse
      > 3 a : to come up again for consideration b : to come again to mind
      > ------> 4 : to occur again after an interval : occur time after time
      > - re·cur·rence /-'k&r-&n(t)s, -'k&-r&n(t)s/ noun
      >
      > Both entries from Merriam-Webster's Collegiate Dictionary.
      >
      > HTH,
      > -Paul
      >
      >
      >
      >
      > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

      --
      -------------------------------------------------------------
      Tex Texin Director, International Business
      mailto:Texin@... Tel: +1-781-280-4271
      the Progress Company Fax: +1-781-280-4655
      -------------------------------------------------------------
    • P A Hill & E V Goodall
      ... The endpoints are when the time on the wall reads what one of the end points says. As I said the end point are dimensionless, including each or or not
      Message 2 of 4 , Sep 19, 2001
        Tex Texin wrote:
        >
        > Paul,
        >
        > Thanks for the reply.
        >
        > 1) Using 3.17, I would presume the endpoints are not included (because
        > it says "between").

        The endpoints are when the time on the wall reads what one of
        the end points says. As I said the end point are dimensionless,
        including each or or not does nothing to the length of a duration.

        As I quoted a time-point is dimensionless another more common term is
        'instance' or 'moment in time', but it is also the name of
        a period of time. The time-point or instance occurs at the start of similarly
        named duration. Just like in regular usage. 1 PM is a time-point right
        at the beginning of the 1 o'clock hour, both are known as
        13:00:00.000.

        > However, by 3.26 the endpoints of the interval are separated from the
        > two time-points by the width of a dimensionless instant, so practically,
        > I can include them... (Meaning, I physically start and end the clock
        > concurrent with the time-points, and the next interval starts at the
        > same time the previous one ends.)

        That sounds right.

        > 2) Although a duration is defined by time-points the spec also allows
        > for reduced precision, so I don't see that using the date format alone
        > necessarily implies the start of the day. I take it to mean the
        > precision is reduced and you don't know where in the day the duration
        > begins or ends.

        No, the reduced precision does not imply that the accuracy is reduced.
        Accuracy is not the same as precision, particularly in this case.

        Look at 5.5.3.1
        "c) If the number of years, months, days, hours, minutes or seconds in
        any of these expressions equals zero, the number and the corresponding
        designator may be absent; however, at least one number and its designator
        shall be present. Note that the removal of leading non-zero components
        is not allowed."

        > This is an important distinction. If using reduced precision implied a
        > more granular time point, then it could cause problems for applications
        > that depended on the reduced precision of the value. For example, if I
        > tell you I will send a message on 2001-09-20,

        Your would require a time-point and a duration, i.e. 2001-09-20 with a duration
        of 1 day, or some other duration.

        > and you assume that since
        > no message arrived by midnight 2001-09-20, it is ok to cancel the
        > transaction,

        Since you seem to be specifying a duration you must use one of the various
        formats which are:
        start time/end time
        Pduration
        start time/Pduration <-- this is the one that seems natural enough.
        Pduration/End time

        Another way to specify that final time would be
        Before 2001-09-21

        >and the message shows up some time during 2001-09-20, we
        > would have a dispute...

        Not if the I understood the thing you specified to be a duration which
        if you used an ISO8601 format I'd know it was a duration. 2001-09-20
        is not a ISO8601 duration it is an ISO8601 date.

        The phrase class will end at 10 AM does not suggest that your average college
        student is going to do anything but get up and walk away as soon as
        the time on the clock reads anything after 10:00:00.000.

        For an example how about specifying a century (using the traditional period,
        not the just minted all year digits wrap.

        The 20th Century
        1901-01-01 / P100Y
        1901-01-01 T 00:00:00 / 2001-01-01 T 00:00:00

        The above two are two different ways to specify the same duration.

        The period starts right at the time lapse into the first second of the first
        day of the year and ends when the clock stricks midnight 100 years later.
        The day, hour or minute that comes just after that is not part of the duration.

        I hope that is clearer,
        -Paul
      Your message has been successfully submitted and would be delivered to recipients shortly.