Date:         Mon, 16 Aug 2004 02:49:25 -0400
Reply-To:     Richard Ristow <wrristow@mindspring.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Richard Ristow <wrristow@mindspring.com>
Subject:      Re: expanding length of computed variables
In-Reply-To:  <008001c4818f$9a1aa380$9300a8c0@NOTEBOOK>
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 07:45 PM 8/13/2004, Hector Maletta wrote:

>[You change the number of decimal places displayed] with the FORMAT command.
>
>All this is just for visualization. The internal numeric precision of SPSS
>is always 20 integers and 20 decimals.

That's quite right up to the end, which is worth correcting. The internal
numeric precision of SPSS is 64-bit floating point with 52 significant
bits, equivalent to about 16 decimal digits -- quite a bit, but a good deal
less than 40 ("20 integers and 20 decimals"). Because these are
floating-point numbers, the 16 digits of precision are available for both
very large and very small numbers; there is no limit that the numbers be
between 10**20 and 10**-20, as "20 integers and 20 decimals" suggests.