Some c-f-90 correspondents have pointed out some errors in my last
message. I apologize for my lack of proofreading; I just ran out
of time.
Roger Glover wrote:
<snip>
> However, was somewhat disappointed with the lack of detail about the
^^^^ [1]
> specific problem features of C relative to Ada.
<snip>
> - in some cases it is possible not to avoid declaring a variable
^^^ [2]
> type or function return type, in which case the type is assumed
> to be "int" (integer)
<snip>
> This
> is a valid and important criticism that applies even more so to
> modern Fortran than to archaic K&R C;
^^^^^^ [3]
<snip>
> This criticism also applies to modern Fortran even more strongly
> than to C, ^^^^^^ [3]
[1] "However, was" should be "However, I was"
^
[2] "it is possible not to avoid" should be "it is possible to avoid"
^^^ DROP THE "not"
This actually says the opposite of what I intended to say.
[3] "modern Fortran" should be "current and projected Fortran"
^^^^^^ ^^^^^^^^^^^^^^^^^^^^^
This one is more subtle. Someone pointed out that "modern
Fortran" features *allow* the kind of programming that Dr.
Zeigler seems to desire. My point was more that Dr. Zeigler
found lack of "style policing" to the real problem with K&R
C, which also allows but does not enforce the sort of coding
style he finds to be "safe" (at least, those were the only
specific areas he mentioned).
The current Fortran standard does not enforce any sort of
style, nor does it delete any of the more unsafe features
Dr. Zeigler would have considered "short cuts." Nor does
the projected "2002" standard. Hence, I conclude that Dr.
Zeigler would be even less impressed by "current and
projected Fortran," than he was with decade-old pre-ISO C.
-------- Cray Research --------- Roger Glover
-- A Silicon Graphics Company -- http://home.cray.com/~glover
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|