Walt Farrell wrote:

| Are you perhaps confusing AMODE and RMODE,
| John?

I am, I must suppose, confused about some things at my advanced age;
but that is not one of them.

There are two issues here that it is useful to disentangle.  I execute
code above the bar routinely; and I know of others who do so too.
IBM's strictures are entirely reasonable.  They come down to the
notion that you are on your own if you do something that IBM is not
yet ready to support.

That said, the future of the mainframe, assuming hopefully that it has
one, is above the bar.  Moreover, in my now considerable experience
with AMODE(64) code I have not found, systems services aside, any
interesting differences between the behavior of such code above and
below the bar.

This morning, for example, I tested some extensions to an AMODE(64)
table-search routine.  As is my habit, I tested it for the 2^2 = 4
relevant situations: table and code below the bar, table below the bar
and code above it, table above the bar and code below it, table above
the bar and code above it.  I expected none, and there were no
behavioral differences.

A little more candor about some of these things is now, I think, in
order.  IBM has provided us with what its lawyers call constructive
notice of the things it is not yet prepared to fix if they malfunction
or fail to function at all.  This, as I have already said, is sensible
and appropriate.  There is no need for further assertions that things
that manifestly do work may not or for something less than clarity
about, for example, the fact that AMODE(64) code is faster than
AMODE(31) code.

John Gilmore, Ashland, MA 01721 - USA

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to