HP-12c, November 25, 4046, answer
Message #1 Posted by Don Shepherd on 3 Dec 2010, 7:35 p.m.
In this thread, I asked the question why does the 12c calendar function have an end date of November 25, 4046. I noted that the number of days between the beginning date of October 15, 1582 and the ending date is exactly 899,999, and I wondered why the specified ending date was chosen by the 12c designers. Thomas Klemm thought it might have something to do with the fact that there is uncertaintly whether the year 4000 will be considered a leap year, so perhaps the designers ended the date range at that point because of that. Gerson Barbosa found a 1977 Hewlett-Packard Journal article by Roy Martin, one of the designers of the HP-92 financial calculator, who stated that the ending date was determined by internal register limitations. Thomas Klemm did some analysis of the actual 12c code and did not see the need for this limitation.
Gerson Barbosa found the current address of Roy Martin and I wrote him a letter asking for clarification on this issue. I received his very nice reply today in the mail. It turns out that Thomas Klemm was correct. Mr. Martin stated that they realized that calendar adjustments needed but not yet decided could cause actual calendar dates to deviate from the calculated results in the year 4000 timeframe, hence the choice of the 900,000 limit on the calendar function on the HP-92 financial calculator (and subsequent HP-12c).
Per Mr. Martin: The claim (in the HP Journal article) that the date was determined by internal register limitations and not any special knowledge of the future was tongue in cheek.
To make a long story short, the designers of the HP-92 limited the calendar days function to around the year 4000 (as opposed to year 9999, for example) because they couldn't know if it would one day be considered a leap year, as Thomas Klemm had proposed.
It is great that at least one vintage HP calculator designer is still around to answer such questions.
|
Go back to the main exhibit hall