Shachar Shemesh <shachar@shemesh.biz> writes:
> Tom Lane wrote:
>> I seem to recall also that there are machines that
>> treat the exponent as power-of-16 not power-of-2.
> I'm pretty sure I don't understand this. Maybe I misunderstood the
> format, but wouldn't that actually lose you precision with, at most,
> marginal gain in range?
I didn't say it was *better*, only that it was *different*.
Some marginal googling suggests that only IBM made this choice, and that
they now (thirty-some years later) concede the IEEE format is better.
Still, if we're running on s/390 hardware ... or ARM, or whatever ...
it behooves us not to lose data. I'm fine with doing byte-swapping to
present a uniform face for platforms that are at bottom implementing the
same FP format; but I've got a problem with trying to force
noncompatible platforms into that format.
regards, tom lane