Agreed. it is a very serious question.
I don't know exactly what a line of code means either, having programmed
in assembler (consider a single instruction vs. a macro, where both are
a single line of source code). And oh yes, I have also programmed in
PL/1 and then created some very complex programs and algorithms in a
single line of APL, where you can do an awful lot of complicated
computing in a string of 130 characters (as on a 2741).
The real issue of size seems to depend on what you want to really
measure. For example, we have used "lines of code" for a long time to
estimate the time required to develop programs or whole systems
(remember "The Mythical Man-month"?). Resulting byte count or machine
instructions don't mean as much either, as really sloppy code can be
substantially larger than it's most efficient equivalent, so which is
better? Realize that even the most efficient assembler program can be
made less effective by changes in the architecture, such as pipe-lining
or other similar advances.
A few years ago, we pretty much dropped the concept of MIPS, changing it
from Millions of Instructions Per Second to Meaningless Indicator of
Processor Speed. Simply compare the accomplishment of a single line of
assembler, such as comparing LR to CFC (Compare and Form Codeword) or
UPT (UPdate Tree). It becomes obvious that a lot more LRs (or similar RR
instructions) can be accomplished per second than the number of CFCs or
UPTs in the same interval of time on the same machine.
So why size matters, except in sporting events such as sumo wrestling or
in other topics into which I would prefer not to go here, I don't have
an answer either, and, for that matter, I'm not sure I really care,
except when I have to get a bigger PC to update my Windoze system to the
next version :-).
Mike Myers
Senior z/OS systems programmer and consultant
Mentor Services Corporation
on the saOn 09/06/2016 02:35 PM, R.S. wrote:
W dniu 2016-09-06 o 13:39, Elardus Engelbrecht pisze:
John Eells wrote:
What's a "line of code"? That's a serious question.
Indeed!
This is one of completely misleading measures very popular in
pseudo-technical environments, like airline magazine articles, or used
by "managers".
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
