Re: times, dates, images, and S-W

[cbmvax!uunet!math.ucla.edu!jimc]

I also am a little slow getting caught up.  On Arthur Protin's use of
yymmdd date format, there is in fact an ISO standard prescribing this
order.  I don't have the number right handy but it's in the library and
I could re-check features if anyone's interested.  Here's my
recollection from several years ago: 

The date consists of fields of digits separated by anything else.  The
standard wasn't exactly clear what was or was not allowed as
separators. The first number is a year.  I don't remember if there was
a prescription for what to do with two-digit years.  Next is a month
and last is a day. Time-of-day was not addressed.  The Gregorian
calendar is (tacitly?) assumed.  

In -gua!spi I have a "vector" construct in which X2, X3, ... (as many
as needed) are expressions and X1 is an ordered list with these
members. This fits right in with dates; X1 (one of these vectors) is
the date of event X2 starting with unit X3 (default years) with
calendar style X4.  A transitive compound with "hours" gives you
time-of-day.  

I wonder if time-of-day or date should have the preference?  In a
language with a very mechanistic style you must choose one; you can't
rely on semantics to suggest which is meant, but you don't want
separate primitive words for both concepts.  

> Now if you want to redefine all our clocks to fill a day with 100,000
> units approximately equal to 0.864 seconds ...

In the late 50's, Donald Knuth (yes, THE Donald Knuth) published a piece
entitled "The Potrzebie System of Units" in Mad Magazine, in which he
did just that.  If I remember right, 1E-5 * day was called a "martin".
The article helped me greatly to understand unit systems and their
metrological considerations -- much more than the explanations I was
getting in school.  

		-- jimc