[J3] [BULK] SCALE(x,i) with complex x

[Malcolm Cohen]

Hi Steven,

> From a mathematical stand-point, there is an implicit set of parentheses
in real*complex.
> r*z = r*(x+Iy) = r*x + Ir*y,  (I being sqrt(-1))

As a (pure) mathematician, I disagree. You are talking about one method of
representing a complex number. Other methods exist, but more importantly,
complex numbers are just numbers.

There are no parentheses in a complex number. None.

That means that 10.5.1.2.4, which in any case refers to parentheses *in the
program source code*, does not have anything to say about what algorithm a
compiler may use for complex arithmetic.

> I suppose that one could argue [that the program is not conforming, and
the compiler does not support IEEE arithmetic]

That is absolutely not what I was arguing at all.

> I concede, and won't bother J3 with other suggestions.

How can you concede over something (gfortran being pessimal) that does not
affect your suggestion?

Suggest away!

All suggestions need critical examination. It's the committee as a whole
that decides what goes in, not me. I'm just giving my opinion here, which is
that I did not find the use-case compelling. Others frequently have
different views.

Cheers,
-- 
..............Malcolm Cohen, NAG Oxford/Tokyo.