[zokier]

I don't completely agree; when using a year as base unit, it makes somewhat sense both directions. Deka- and hectoyears are equal to common decades and centuries, and megayears would be relevant for geology and astronomy etc. In other direction deci- and centiyears could be sensible, equaling a decimal "month" and "week" respectively.

Playing devils advocate here, the problem with this proposal is that it didn't go far enough with its decimalization. A "better" (more decimal) approach would be to divide year to 100 "weeks" (10 "months" with 10 "weeks" each), and have each "week" vary in number of days. This would open up to fun ways to notate it, for example now (2023-12-17T21:03:31Z) could be "2023.95/2.87744Z", i.e. "week" 95 (or month 9 week 5) of 2023, day 2 of the "week", and 877.44 millidays ("minutes") since midnight. It's only slight abuse of decimal notation :)