It reads in to Thunderbird OK, but the result is less pretty than
Eliot's effort. Eliot's version appears to lay out the multiple events
in a partcular timeslot evenly across the available space, whereas
Bill's results in different sized blocks and in some cases some of the
events appear to be hiding others.
This may be because mine treats two sequential 1-hour meetings as a
2-hour meeting, and Eliot's treats them as 2 1-hour meetings. See
l3vpn on Monday and mip6 on Tuesday evening. I originally wrote the
agenda parser that's the basis of this .ics output when we switched
to 1-hour meetings on Tuesdays because I could never tell whether
a given meeting was one or 2 hours long.
I have yet to find a client that has a good variable-length overlapping
event rendering algorithm, especially with 8-9 events in the overlap.
Apple's iCal usually does really well when there are 3 or 4.
Ietf mailing list