No, they just define what a date (either calendar or ordinal) is.
Without saying what one means or how to use it.
But also look at 3.17 and 3.26 -- 3.26 defines a time-point, and 3.17
(and 126.96.36.199) specifies that the values in a start/end interval are
time-points (I think I mistyped it as 188.8.131.52 in my earlier post,
And, of course, 5.5.5.d
I guess it boils down to this: a time interval with start and end
time-points with "reduced precision" (i.e. no time of day specified)
can only be used "by mutual agreement". It is perfectly valid and
possibly probable that the agreement will be that the start value
would be "T000000" and the end value be "T240000", as you argue. But
it doesn't have to be, another partnership could use "T090000"
and "T170000" respectively as their time of day values.
I myself would demand with any and all who wish to interchange with
me, that "reduced precision" _never_ be used. The computer doesn't
care and I think network speeds are sufficiently high that sending
those "extra" bytes won't matter a gnat's arse.
Allow me to point out that I doubt your nominal meeting will convene
at midnight one one day and run through to midnight on another day.
It seems likely you would hold it from 090000 to 170000 on each day,
meaning you would need to send a time interval format for _each_ of
the meeting periods.
I think what you are talking about is the more informal "human
interchange", saying, "we're going to meet from 9:00 to 5:00 on the
10th through the 11th". For which ISO8601 is probably inappropriate.