Path: utzoo!attcan!uunet!mcsun!hp4nl!charon!cwi.nl
From: jurjen@cwi.nl (Jurjen NE Bos)
Newsgroups: comp.sys.handhelds
Subject: Re: Spectral test on HP28/48 random generator
Message-ID: <2079@charon.cwi.nl>
Date: 6 Sep 90 14:34:23 GMT
References: <2056@charon.cwi.nl> <1990Sep5.041507.13226@laguna.ccsf.caltech.edu>
Sender: news@cwi.nl
Organization: STORC, Veldhoven
Lines: 18

Mark Adler, thanks for your reply.  It doesn't surprise me the least that 
they used a multiplier that passes with "flying colors".  Isn't that what
every HP buyer expects them to do?  We want _the best_. We buy HP.

Also, you write:
> I don't know how the seed is set on the HP, so I don't know if the seed
> condition (not a mutliple of 2 or 5) is met.  If the seed condition is met,
> then 5x10^13 is a more than adequate period, since only 12 of the digits
> become part of the actual random number.

The seed condition is met by forcing the lst digit to a one.  To be exact,
the seed is set from the real number given as first the 12 digits of the
mantissa, then the last two digits of the exponent, then a 1.
--
|                 | "Never imagine yourself not to be otherwise than what |
| Jurjen N.E. Bos | it might appear to others that what you were or might |
|                 | have been was not otherwise than what you had been    |
|  jurjen@cwi.nl  | would have appeared to them to be otherwise."         |