[J3] [EXTERNAL] Private enumerators

[Van Snyder]

On Sun, 2019-10-27 at 09:08 +0900, Malcolm Cohen via J3 wrote:
> Doing a hack job is worse than doing nothing.

I wholeheartedly agree with this.  I have been involved with Fortran
standardization since about 1985, and a member of J3 since 1997.

I have seen many half-measures done, either because there was
controversy over trivial points of syntax, or we just didn't want to
invest the effort to do it right the first time.

Doing it wrong the first time effectively casts an eternal kludge in
concrete.

The example you have heard me use frequently is the definition that a
disassociated pointer corresponding to an optional nonpointer dummy
argument is absent.  This was done precisely because we refused to do a
conditional form, but we agreed there was a combinatorial explosion
problem.  Before the kludge arrived on the work list, the conditional
form was on the work list.  It was removed because of trivial
disagreements over syntax, and an aversion to doing the work necessary
to do it right.