Re: Lost in free format conversion.

On 25 Jan 2013 12:05, Henrik Rützou wrote:

Come come

the difference in a positive or negative value is in the last byte
that always are there signed or packed

binary 1111xxxx is positive giving x'Fx' a positive value
binary 1101xxxx is negative giving x'Dx' a negative value

"Come come"? Is there something that was offered in the most recent reply [quoted above] that would contradict anything that I stated in my earlier reply [quoted below]?

Missing from the above reply was mention of each of:
binary 1010xxxx is positive giving x'Ax' a positive value
binary 1011xxxx is negative giving x'Bx' a negative value
binary 1100xxxx is positive giving x'Cx' a positive value
binary 1110xxxx is positive giving x'Ex' a positive value

Regards, Chuck

On Fri, Jan 25, 2013 at 8:56 PM, CRPence<CRPbottle@xxxxxxxxx> wrote:

On 25 Jan 2013 03:05, whatt sson wrote:

<<SNIP>>in EBCDIC a byte representing a character "1" is the same
as a byte representing a zoned numeric 1, which is hex F1. So
MOVE works coincidentally, with alpha fields and zoned numeric
fields as long as the number is positive. The "F" in F1 indicates
it's "unsigned", a "C" indicates it's positive and "D" indicates
it's negative. The "sign" in a zoned field is represented in the
last byte (least significant). That's why you get the letter "O",
which is hex D6, representing the digit 6 in a negative number.

FWiW: While that effectively describes the Binary Coded Decimal
(BCD) for Zoned Decimal on the IBM Mainframe, the midrange chose to
eliminate the concept of unsigned digits; choosing instead to have
its preferred _positive_ as the hex digit "F" and its preferred
_negative_ as the hex digit "D". The only consequence that I am
aware of, is that the one-digit decimal number of zero is a
/positive/ value represented by x'F0'. And while the 0xD and 0xF
are preferred signs, all of the hex digits 0xA to 0xF are valid
representation for a sign.