The construct

| . . . on machines supporting a negative floating point zero.

was, not very long ago, unambiguous.  zArchitecture machines now
support three different kinds of floating-point arithmetic (with much
sharing of implementation machinery), viz.,

o IBM hexadecimal floating-point, HFP,

o ANSI binary floating-point, BFP, and

o ANSI decimal floating-point, DFP.

Of these the first two, HFP and BFP, make zeros positive; but the
third, DFP, supports both positive and negative zeros.

There are arguments pro and con.  The mathematical  rationale for a
signed zero is persuasive.  The anecdotal argument against it was once
familiar.  The old among you have heard [too many] war stories about
the Widget Mfg Company customer who was dunned over and over again
because he had a positive debit balance of zero.  Regrettably, these
stories have some basis in fact.  Care, which is always in short
supply, is required to avoid such programming gaffes.

John Gilmore, Ashland, MA 01721 - USA

Reply via email to