Stan Sieler wrote:
> Ok...the hazard of being too old :(
>
> By "Does anyone use Chronos" I meant the old, old CHRONOS *intrinsic*
> from the SERIES I days... the one that Bob Green wrote, which returned
> the date/time in a 48-bit floating point.  Later models of the 3000
> dropped 3-word (48-bit) reals in favor of 2-word & 4-word reals,
> and CHRONOS was dropped from MPE.

Nobody has used CHRONOS in two decades unless they had a pre-Series-II.
I am old enough to have presumed a 3-word entity, but speculate as to
the "3-word real" value.  As I recall, CHRONOS simply returned BY VALUE
the 16-bit CALENDAR value in the high-order word, and the 32-bit CLOCK
value as the low-order words.  As such, it wasn't a "real" floating
point value, just a way to return a timestamp (by value) to the stack.
Many applications (even today) store a "timestamp" as a 48-bit value in
concatenated CALENDAR/CLOCK format.

But I could be wrong since I never got to actually "call" CHRONOS.  When
we bought our first 3000 (in 1976) we had dial-up access to a demo
account on a 3000 in the Atlanta HP office we used for "migration" and
preliminary development.  I was tagged to be the one to develop some SPL
SL procedures to support our COBOL applications, and thus had a
handed-down set of SPL/Intrinsic manuals (from the old CX days).  My
first intrinsic attempt was [you guessed it] CHRONOS.  It never worked,
giving me an UNRESOLVED EXTERNAL at runtime.  Since I didn't know beans
about the 3000 then, I spent a couple of days trying to "fix" it before
finding out that our demo account was on a series II (no CHRONOS).

Jeff Kell <[log in to unmask]>