Dave Cole noted on Monday, March 18, 2013 2:34 PM re:
At 3/18/2013 11:18 AM, Andreas F. Geissbuehler wrote:
- IBM opcodes will remain [A-Z]
Hmmmm... TRAP2 and TRAP4 come to mind.

Schwarz, Barry asked on Tuesday, March 19, 2013 2:42 PM re:
My 2 bits:
- IBM opcodes will remain [A-Z]

Did you mean first letter or entire name?  Many macro names contain
digits and at least one contains lower case letters.

Right you are ! So, let's add [0-9] and [a-z] to [A-Z] for Opcodes.

My intent was to advocate that we could easily correct past omissions
if IBM accepts that all (past and future) *opcodes* are 1-(5..6)
alphanumeric charaters [A-Z] [a-z] [0-9], the first one always an alpha.

That would cover us users and ISVs part way. We still risk to encounter
the same macro name in two or more in-house, ISV or IBM maclibs,
even if we lace our macro names with [@#$] characters like
DC$USING, CR@EX, TR#TBL, etc.

Now, down the road, if IBM were to implement a new concept, they
could open up a previously illegal character for opcodes like the
hyphen to clean up the opcode mess and address the issue raised by...
Jonathan Scott,  Sunday, March 17, 2013 12:18 PM
for example we have "Jump Long Unconditional" (JLU)
because JL was already taken.
J-L    jump LOW
JL-U    jump long (unconditional)
J-NE
JL-NZ
BR-O
or, just append the CC-mask:
AL-0110
BCR-1100

Another partly implemented one is the colon ':' where I'd like
a colon-suffix like MVCL: to mean MVCL:MAC, letting me
write by own instruction syntax like
     BC:   NZ,LOOP1
     BC:   UF,UnderFlow
or just generally, for clarity's sake, mark my macro-'opcodes'.

Lastly, if it were so bad that IBM always names its new opcodes
and macros after some universally used in-house or ISV ones
then KRCB:U could be another future expansion to mean
"macro KRCB from SYSLIBU".

I believe it isn't too late to tighten up now what our (grand-)
fathers would have carved into granite had they foreseen their
ASM would morph 50 years later into what HLASM would be
if its developpers had sufficient funding...

Andreas F. Geissbuehler
http://www.afgc-inc.com/

Reply via email to