Regarding the proposed new date intrinsics, I do second
Stan's suggestion to use HPE_STATUS (i.e. status,subsys)
for the error parameter like all/most of the HPsomething
intrinsics do.

For me the HPDATEADD functionality is somewhat confusing.
What does it mean to add 1 month or one year? How many
days are that? What is the result of 30-Jan-1996 plus one
month for example?

How about having a date-to-"linear-date" conversion similar
to the date-to-integer and integer-to-date $control post85
functions in COBOL? Wouldn't that provide the kind of basic
service for date-calculations i.e. date plus n days/weeks
or number of days between to dates? The "origin" of this
"linear date axis" is pretty irrelevant for the offsetting
or subtracting operations, by the way.

I also second the suggestion to include a function for
validating dates as this seems a pretty frequently needed
operation.

Bahut shukriya to Kishore for posting on HP3000-L.

Lars :-)