3 ms·
It seems to me, you'd need to redefine the second such that there is a nice round number of them in a day. Say 100,000 to the day. They you could subdivide into
by cynwoody 13y ago
It seems to me, you'd need to redefine the second such that there is a nice round number of them in a day. Say 100,000 to the day. They you could subdivide into hectoseconds (new minute, 86.4 old seconds), kiloseconds (14.4 old minutes), myriaseconds (tenth of a day, 8640 old seconds). Noon would fall on a nice even boundary.
(Of course, this ignores the fact that the day length is not constant, but we have that problem with the present system, too.)
The myriasecond would be the new, lower-resolution hour, much as the Celsius degree is lower-resolution than the Fahrenheit. This would be mitigated by writing decimal times (e.g., 3.75 myriaseconds on the clock would correspond to 9 am).
You'd also need to come up with less clunky names than myriasecond, I would think.
Edit: Why it'll never happen: Redefining the second would screw up too many things other than time of day references. E.g., frequencies and other physical measurements based on the second.