>
> > > and "selected_real_kind(12)" might do it better.
And I said
> > If you think it obvious that this means 64-bits, be my guest.
Robin replied
>If you know of a better way, then use it.
I don't. That was sort of the whole point - that I don't know of
anything I consider to be a good way. That's why I'm proposing
a new feature.
>Most folks think in terns of the number of decimal digits required.
>They don't think of 64 (bears no relation to the actual precision).
My observation of "most folks" does not support this claim. In my
experience, most folk think in terms of wanting the 32- or 64-bit real.
Perhaps from an abstract viewpoint this might be considered strange,
I'd agree. But in my actual observation of people using the language,
that's how the majority of them think; they just find it confusing and
complicated when asked to request sizes in terms of decimal precision.
Sometimes they just refuse to do it, preferring nonstandard forms
such as real*8. Other times they try to translate their request for
64 bits into precision and get it wrong, asking for slightly too much or
too little precision to get the 64 bits that they wanted.
I realize that the 64 bears little relationship to the actual precision and
might abstractly be thought a strange thing to request. I'm still going
to claim that's what "most folk" do, whether it makes sense or not. I'm
one of those "most folk" by the way.
|