On Mar 11, 2004, at 9:48 AM, Renchi Raju wrote:
> maybe this has been suggested (and rejected) before. but how about
> having
> something similar to implicit statements.
>
> for eg:
>
> implicit real_kind(selected_real_kind(12,30))
Yes. With trivial spelling variations, that's among the first of
the ideas mentioned. Alas, how to spell the implicit statement is
by far the smallest of the problems.
If this were the only issue, it would have been in the standard
long ago. It was certainly asked for long ago. Indeed, it was
*ACCEPTED* as a proposal for one of the standards (my time sense
gets off enough that I forget whether it was f95 or f2003; it was
quite a few years ago, but then f2003 work was started quite a
few years ago.) The reason it later got dropped was not because
people didn't like the idea, but because it proved difficult to
settle on and precisely specify all the cases.
How to specify exactly what it does in all cases is much harder.
And avoiding big implementation burdens and big opportunities for
undetected errors are also hard issues.
Note that the vendor implementations vary quite a lot on the fine
details of exactly what happens in all cases.
P.S. No, I don't have the time to explain all the complications.
I don't even remember them all and I'm too busy today. And one of
the problems was that overlooked things kept coming up - we might
well not have yet uncovered all the issues before giving up.
You'll just have to trust me (or not as the case may be) when I
say that it isn't as simple as you might think on the surface.
--
Richard Maine | Good judgment comes from experience;
[log in to unmask] | experience comes from bad judgment.
| -- Mark Twain
|