Re: "Classical Fortran" Second Edition
- From: Terence <tbwright@xxxxxxxxx>
- Date: Sat, 18 Apr 2009 16:17:25 -0700 (PDT)
In article <VXlGl.12383$jZ1.2409@xxxxxxxxxxxxxxxxxxxx>,
Gary Scott <garylscott@xxxxxxxxxxxxx> wrote:
Luka Djigas wrote:
** over ^ has an advantage in that some of non-english keyboard
layouts don't have that symbol, so the users have to switch to EN then
yes ^ is a terrible choice because so many keyboards don't have that
symbol. On some, its a .not. symbol above the 6 key.
I remember one language (which I never actually coded in) that used
the proper character for that: vertical arrow :-) And, yes, the paper
tape punch had that character on it (a flexowriter?)
1) whatever "The Committee" chooses among symbols to use in Fortran,
should be from the 7-bit ascii set to ensure that practically all
keyboards wil have them..
2) I don't think I'll use them myself, but I will accept that ">"
">=", "<", "<=" are reasonable alternative symbols for
But I really hate the idea of "==" for ".EQ." when "=" should serve.
There can't be any parsing confusion.
The "==" seems to mean means 'equal or equal', a tautology.
3) And nothing looks reasonable to replace ".NOT." ,
I remember we used the (top left) corner symbol in IBM, which I
don't have on my keyboard, for ".NOT."
And don't some languages use "!" for this? Which of course won't mesh
with this being used for a comment in ortran. How about "~" instead
for ".NOT."? Or is it too easily misread as a minus sign?
Robin mentioned the arithmetic IF.
I agree it was always a bit clumsy in that a large proportion of tests
were two-way and not three-way, but so useful when you needed that
separate "what if zero?" special case, and ended up with two tests and
branches where there was only one before.
I DO use the logical IF tests myself, and never the arithmetic IF,
even if I write a tiny bit more code. But I insist that the computed
branch is a more readable structure that the huge CASE clauses, which
seems to have been invented just to remove labels.