Re: Let's not BE NP-hard, shall we...?
On Aug 5 2011, Ian McIntosh wrote:
I think I understood your point about explicitly versus effectively, and we
mostly agree. I just don't agree with Dan that order of evaluation is just
"effectively not specified" in Fortran, or with you that the order is
always "explicitly unspecified". I believe the truth is between.
I apologise if I confused people by implying that it is always unspecified.
As you have described, it isn't. Fortran is explicit about what it
specifies and what it does not, but that is not simple and its implications
are not obvious.
The main difference is between it and standards like C, where both the
specification that the evaluation order follows the syntax order and that
certain expression evaluations are unspecified exist only by implication.
What is worse, is that there is another implication that the former is
NOT required! I.e. you have to read between the lines, not everyone
reads them the same way, and (even in WG14 itself) the consensus has
varied over time. That is not nice :-(
Regrettably, this IS relevant to this debate, as one has to consider not
just the actual wording of the C standard, but its present and future
interpretations by important C-using communities.
But I think that the point raised by Vladik Kreinovich is the killer.
Attempting the mathematically impossible just isn't going to work.
Regards,
Nick Maclaren.