Re: Integer Coersion

On 2007-02-07 09:38:24 -0400, mecej4 <mecej4@xxxxxxxxxxxxx> said:

PJH wrote:
<SNIP>
> ps
Someone around here is going to get hot under the collar with you for saying
that INTEGER(1) implies 1 byte, rather than INTEGER*1. I'm sure there
actully are some compilers out there that deliberately don't have a
KIND=BYTE equivalence.

David Frank wrote
<SNIP>

I am not sure about "deliberately", but Salford FTN95, for example, considers the kind numbers to be ordinal, giving 1,2,3,4 for INTEGER*1,2,4,8 and kind=1,2,3 for 32, 64 and 80 bit floats.

After all, the English word "kind" suggests enumeration rather than measurement. On the other hand, if each increase in the kind number suggested a doubling of the byte-count of the floating-point word, the 80-bit float would be "odd".

-- mecej4

It has been repeatedly observed in the past that byte size is not a unique
specification of a floating point format. There can be an exchange between
the lengths of the exponent and significand fields.

There have been major vendors with popular computers lines for which this
was true. Consider the various VAXen.



.