On Monday, January 23 2006, "[EMAIL PROTECTED]" wrote to "[email protected]" saying:
> Hi, > I have a question regarding the interpretation of $,1rx(Bduration$,1ry(B > attribute > value in 'time$,1ry(B node as mentioned in RFC 3880 (CPL). > RFC 3880 - Section 4.4 - Time Switches > "For a recurring interval, the "duration" parameter MUST be small > enough such that subsequent intervals do not overlap." > In the RFC, there is no further discussion on this statement. Please let me > know which of the following interpretation is correct. The requirement is intended to ensure that by the algorithm given in the appendix, there is only one possible candidate start time for any given time, while still allowing maximum flexibility in the permitted recurrence rules. Reducing this rule to a specific constraint is rather difficult, which is why the rule is stated in this somewhat roundabout way. > 1) 'duration' must be smaller than the 'freq * interval' value (byxxx is > not considered) > <time dtstart="20050301T093000" duration="P700D" freq="yearly" interval="2" > bymonth="3,4"> > Here 'duration' value should be smaller than (365 * 2 days) This is not valid -- 2005-04-02 00:00:00 appears in two different intervals of the recurrence. > 2) 'duration' must be smaller than the minimum unit (taking byxxx into > consideration) > <time dtstart="20050301T093000" duration="P28D" freq="yearly" interval="2" > bymonth="3,4"> > Here 'duration' value should be smaller than 28 days (maximum difference > between bymonth attribute values - assuming a month difference = 28 days) This is a sufficient, but not necessary, constraint. In particular, for this rule, P31D would be a valid duration, because March has 31 days; and if you had a bymonth="3,9" parameter, P100D would be valid. However, for bymonth="2,3", P28D would be the maximum valid duration. -- Jonathan Lennox [EMAIL PROTECTED]
_______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
