RE: Strange behavior w/%editc()

["Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>]

> From: Jim Langston
>
> Actually, I know which fields are from database files by their names.  So
> far every application I've worked on, or written, for the AS/400
> so far have
> had naming schemes to distinguish field variables from program variables.

Thank you, Jim.  You've made my point quite eloquently.

If your naming conventions are enough to tell you the contents of the
fields, then it follows that you don't need the redundant information that
the %editc and %dec syntax provides.

My point is that counting on the syntax of your statements for information
about your variables is perhaps not as important as it has been touted in
some of the posts here - certainly not important enough to triple or
quadruple the size of the instruction, adding two levels of indirection.
You should get the information about your variables someplace other than the
data movement instruction, whether it be a MOVE or an eval.  You should make
coding the instructions as easy as possible.

By the way, this argument lessens considerably with object oriented
programming languages.  Since the goal in an object-oriented environment is
to never write the same line of code twice, and every method of every object
is in effect a new opcode, then it makes sense to have long, descriptive
names for everything (not to mention the fact that in an object oriented
environment, you shouldn't use global variables anyway).

RPG, though, as pointed out by Hans, is anything but an OOP.  Since RPG is a
procedural language and you have to code lots of data movement
instructionsw, it just makes sense to make that particular operation as easy
to code as possible.

But hey, I'm not here to convince anyone, I'm just clearing my own thoughts.
Debating the issue with all of you has made it absolutely clear to me that
my views are correct, FOR ME, and that I'll simply avoid /free until such
time as primitive data movement operations are included.

Joe