In article <elkhr8$51h$1$8302bc10@[EMAIL PROTECTED]
>,
"Mike Cowlishaw" <mfc@[EMAIL PROTECTED]
> writes:
|>
|> > And, as I can tell you from 35 years of teaching, consulting and
advising
|> > on this, decimal floating-point WILL NOT HELP with that. What it
will
|> > mean is that they will have completed their programming course before
|> > they hit problems caused by the floating-point, and so won't have
anyone
|> > to explain the issues to them.
|>
|> Well, I have only 38 years of the same, and I do not have all the
answers.
|> You do, clearly, so please stop criticising and start being
constructive. I
|> don't think the choice of decimal or binary is going to solve the
latter
|> problem.
Aargh! That is what I have been saying all along!
Look, I have NEVER said that decimal floating-point is a disaster, in
itself, though it is possible that people's belief that it will solve
existing problems may cause such a state. ALL I have been doing is
correcting claims that decimal floating-point will bring advantages
that it won't. It is YOU that have been claiming that it is a great
improvement. I have claimed all along that there is a negligible
difference, numerically or in ease of use, and the other benefits are
fairly equally balanced.
IEEE 754, decimal floating-point and IEEE 754R may have won the politics,
but damned if I am going to let them indulge in technical revisionism
and claim benefits that are provably false.
I promise that, if you don't claim that decimal floating-point will
solve problems that you can't show it will, I won't jump down your
throat. Or, at least, I will apologise when I am unjustified in
doing so ....
Regards,
Nick Maclaren.
|
