John McKown wrote:

<begin extract>
Or, amusingly to me, if you have a MACRO of the same name in your
MACLIB, the macro will be expanded rather than the opcode use. So, if
you wanted to, you could create a lot of macros to implement
instruction-equivalents. Then use of a specific opcode would either
generate the code for the opcode if the MACHINE level is high enough,
or use the macro definition if the opcode does not exist at that
MACHINE level.
</end extract>

This is, I think, an intelligent, imaginative design.  It has served
IBM well in the past.   There have, for example, been machines that
did not implement certain floating-point divides; and in these
circumstances a macro has been supplied to close the gap produced by
such an omissis.

It could also be used by ISVs.  FLOGR is not, for example available on
all of the mainframes currently in use, but its availability changes
the rules for bit-map processing in a fundamental way.  Shipping a
FLOGR macro with a product that notionally uses the FLOGR instruction
would eliminate the need, felt by some ISVs, to avoid it (and other
instructions of that ilk).

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