non-binary hardware (was: Absolute size of 'short')

[Steve VanDevender]

In article <297 at sdti.UUCP> turner at sdti.UUCP (Prescott K. Turner, Jr.) writes:
>In article <1285 at mcgill-vision.UUCP> mouse at mcgill-vision.UUCP (der Mouse)
>writes:
>>By the way, does anyone know of a non-mechanical digital calculator or
>>computer that isn't essentially binary?
>
>A couple of years ago I attended a presentation by Prof. William Kahan 
>of U.C. Berkeley.  He had an HP calculator with him which as I recall
>adhered to the IEEE 854 standard for floating point using _decimal_.
>An example of decimal integers would be better in some ways, but given
>the stringent rounding requirements of the IEEE floating point standards,
>this calculator's essence could only be called decimal.
>

All HP calculators that I've ever seen use BCD floating point (ones
that I've used are the 11C and 41C, but I've played with the 15C, 34C
and 28C).  Most older HP calculators use a 7-byte BCD floating point
format with a 10 digit mantissa and 2 digit exponent, with single
digits holding the mantissa and exponent signs, and what is essentially
a bias 1000 exponent (999 represents an exponent of -1, 901 is -99, and
001-099 are positive exponents).  The 28C and 28S (and presumably the 18C
and 17S and so on) use a 12 byte mantissa and 3 byte exponent which
can range from -499 to +499.  HOWEVER, the CPUs of all these machines are
essentially binary, in that they can handle non-BCD quantities and store
data in bits--not in trinary or whatever.





-- 
Steve VanDevender	uoregon!drizzle!stevev	stevev at oregon1.BITNET
"Bipedalism--an unrecognized disease affecting over 99% of the population.
Symptoms include lack of traffic sense, slow rate of travel, and the
classic, easily recognized behavior known as walking."