Eric Sand writes:

> In response to Wirt Atmar's observation that one(1) VAR would be
>  adaquate
>  to represent the concept of TIME (CCYYMMDDHHMMSSMMM) I totally
>  agree and would further sugest that to pay this concept its'
>  appropriate respect
>  I suggest that the VAR be named HPCHRONOS in deference to one of the
>  very first intrinsics written for the HP3000.

Let me say that I very much appreciate Eric's seconding of the format that
Jeff Kell and I proposed. Not surprisingly, I still agree with the content of
the posting. But actually far more interesting than the TIME VARIABLE itself
is the phenomenon is VARIABLE TIME.

I didn't post that posting today. In fact, I didn't even post it this year.
Rather, it was somehow dredged up from the murky depths of ancient HP3000-L
postings and automatically reposted by means, I suppose, of a (insert spooky
music here) "Cyberian Time Warp" (tm, all rights reserved). If you look at
the posting, the date it was posted was November 22, 1996 (more spooky
music).

I was as surprised as anyone to see it again. Maybe it's true: the ill you do
will live on long after you've shuffled off this mortal coil.

Nonetheless, all that having been said, I really do support some form of
STRING variable as a CI variable and callable intrinsic that is a complete
time stamp unto itself that can be parsed as needed. As approximately one
half of the respondants earlier agreed, rollover is a serious problem,
especially with application programs that must create database time stamps or
sequenced events.

Although as a string variable it becomes a little long, think of the
marketing phrase that can go along with it: "From milliseconds to millenia in
one simple variable."

Wirt Atmar