(j3.2006) Lazy evaluation of MERGE

Dan Nagle dannagle
Thu Nov 15 21:24:23 EST 2012


On Nov 15, 2012, at 18:36 , Van Snyder <Van.Snyder at jpl.nasa.gov> wrote:

> On Thu, 2012-11-15 at 16:00 +0900, Malcolm Cohen wrote:
>>> Are there any J3 members who were not in
>>> the /DATA subgroup meeting who care to hear, for example, why lazy
>>> evaluation of MERGE would be desirable?
>> Or indeed, why it would not be desirable.
> That's why we pondered requirements in plenary, not subgroup, in 2004.
>> The committee has debated this before, and it was turned down.

> So, that's why I believe lazy evaluation of MERGE is desirable, even
> though it's an inferior alternative to the correct solution.  Let's hear
> a reasoned argument why it would be undesirable.

Lazy evaluation is discussed in Arjen Markus' book, and is
an example and an exercise for the students in my advanced
Fortran workshops.

Perhaps if we all had a bit more experience with Fortran 2008,
or even Fortran 2003, we would find it easier to solve these conundrums.


Dan Nagle

More information about the J3 mailing list