Martijn Tonies wrote:

>>The first longword is the number of days since 17 November 1858.
>>(That's a long story).
>>
>>
>Care to do a write up and add it to the History part at cvalde.com?
>I'm curious :)
>
>

The VAX was the first DEC machine (to my knowledge) that had "fancy"
datatypes - it was a remarkable CISC (complex instruction set computer)
architecture. Sadly, I was part of the group that led it down that path
- though far to junior to have contributed. The COBOL compiler group
pressured the Star group to include instructions for substring and a
half a dozen other things. Anyway, date was one of the special
datatypes. So they needed a base date and 17 November 1858 was chosen
because it was the day time, as we know it, began.

Some people said it was the date that the Smithsonian exposed the first
photographic plates to the stars, which established a time standard.
Others, who'd already debunked that story, said it was the date that the
Naval Observatory in Greenwich exposed teh first photographic plate to
the stars. That also turned out to be untrue. Keven G. Barkes offers
an explanation here: http://www.kgb.com/dcl/199111.txt. Here's another
explanation http://parris.josh.com.au/humour/work/17Nov1858.html.
Well, you can google "November 17, 1858" as well as I can.

InterBase was intended to be - and was - a transparent replacement for
Rdb/VMS. Change a logical name from Rdb to ISC and off you went. Never
did us much good. We kept the base date but the actual VMS datatype
used 100 nanosecond units since the base date, expressed as a quadword.
At the time, quadwords were rare, so Jim implemented the InterBase date
as two discrete longwords. The Rdb bridge handled the conversion, IIRC.

Regards,


Ann