On Mar 16, 2004, at 12:22 AM, Brent Pinkney wrote:

> A big part of the Chronology package was that DateAndTime instances 
> increase monotnoically - i.e
>
> 	(DateAndTime now) = (DateAndTime now)	=> false.

Sorry, but I'm missing the point of that requirement.  We have a 
millisecond resolution clock.  If I ask for "DateAndTime now" twice 
within one ms, why should I get two different values?  Given that they 
would have to be made different artificially (by incrementing a counter 
or something), how is that useful?

I don't care if the particular changeset I posted gets used, but I care 
very much that "DateAndTime now" be accurate, much more than I care 
that it be precise.  The current situation (off by an unpredictable 
number of minutes from the system clock) is absolutely unacceptable.  
If there's no way to fix this in a way appropriate to Chronology, then 
Chronology should not be in the image.

Avi