RE: Are date ranges half-open?

[Frank.Dawson@xxxxxxxxx]

> This should be explicitly stated, but I would think
> that Time spans should be half-open, and that this should
> be stated in the spec. Otherwise, a one hour event starting
> at T150000 would end at T155959, which looks to me like an
> interoperability nightmare.
> 
> It would seem logical that date spans should be consistent
> with time spans, but either way it should be explicitly
> stated.
> 

I may need a cafeine break, but I was thinking that I was applying the
calendar and time scales for duration in a consistent way. You add the
scalar of the duration to the start. 

The issue was that some folks were converting the calendar scale into a time
scale. In such mixed scale cases (i.e., DTSTART in date-time...effectively a
time scale and the end/DURATION in calendar scale), I would have used the
date-time to determine the calendar scale and then apply the duration to
that.

But enough...

-- Frank