Chris Craddock wrote:
<snippet begins>
. . . However, if you're trying to be more general (and reliable) way of doing
this [is] to provide those callers with different interface names . . .
</snippet ends>
and I, as often, agree with him. This time, however, his language strikes me
as unnecessarily gnomic.
Although early versions of C included 'entry' among its reserved words, this
hint was never taken; and in time multiple entry points in a C function, never
possible, came to be anathematized on the principle that if I can't have it it
must surely be worthless. The 'package' was provided instead.
In a z/OS HLASM environment, however, two or more entry points, minimally one
for AMODE(31) and another for AMODE(64) callers, provide the appropriate
mechanism for addressing this issue.
Why have the people who have had their say in this matter failed to note this?
Speculation of this sort is mostly idle, but I suspect that naif notions of
what is and is not 'structured' or even the fear that a proposal might be
dubbed 'anarchic' played some role in this systematic omissis.
John Gilmore
