Re: 2017-03-06- UTC, TImeZone, DayLight Saving Shifts, Enconding

On 19.03.2017 18:23, Cyrus Daboo wrote:
> Hi Joe,
>
> --On March 19, 2017 at 8:45:48 AM -0700 Joe Touch <touch@isi.edu> wrote:
>
>>> look at this image:
>>> https://cdn.pbrd.co/images/LkqU6dJCk.png
>>> there you see the end of the "appointment" is the beginning of the next
>>> day which is invalid ...
>>
>> That is correct if you interpret 'end' as the time when something 
>> ceases,
>> not the last instant inside the interval.
>>
>> But this is far out of scope for the ietf.
>>
>
> Actually the IETF does specify that in RFC5545 (iCalendar). iCalendar 
> defines the time range <start, end> as start being "inclusive" and 
> "end" being exclusive.
>
> However, for all-day events, showing a 1 day event on Monday as 
> "Monday -> Tuesday" is certainly "odd" as most users don't appreciate 
> the difference between "inclusive" and "exclusive". 
Exact this is the bug, because some logic uses this and follows invalid 
things ...
when the client allows to define, only to show events going on the 
actual day or later, such events are shown, too ...

by the way, the RFC5545 is bullshit, because nobody interprets the following
"working days from monday to friday"
like this ...
the end is always inclusive ...
and should there be calculated correctly, that's all ...