Re: Gentler Decimal Floating-Point

Nick Maclaren wrote:

In article <1175079126.776658.6270@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"Quadibloc" <jsavard@xxxxxxxxx> writes:

(snip)

|> > Would you like to define what the logical bit operations (and, even more
|> > so, shifts) mean for an all-decimal computer?

|> No such operations exist.
(snip)

Ah. So what would people do if they had requirements that necessitated
the use of bit arrays, complete with shifting?

The Fortran standard seems to allow it. K&R C allowed decimal, but ANSI
removed it. Fortran allows the implementation to specify the number of
available bits for bit operations, which may be less than that available
for numeric operations. Specifically, BIT_SIZE() for bitwise operators,
and RADIX() and DIGITS() for numerical values, either INTEGER or REAL.

There are a few possibilities, some require more work to perform bitwise operations than others.

-- glen

.