Re: Some ILE questions.

There are multiple differences between a subprocedure and a subroutine.
1 - A subprocedure supports local variables. for example you could have X
defined as a 10 character field in the main body and as a 5 digit integer
in a subprocedure. Thus if you use the same subprocedure in multiple
programs you don't have to check the variable names. Not possible with a
subroutine.
2 - A subprocedure supports parameters, and, a return variable. For
example
x = MySubprocedure( a : b: c);
will set x equal to the value returned by the subprocedure MySubprocedure.
Sort of like building your own RPG bif's.
3 - You could copy in a subprocedure using /copy (like a subroutine)
however, many people chose to put the subprocedure into either a separate
module, or even into a service program. In this way it's poorly compared
to a completely separate CALLed program.
4 - Prototyping is a bit of extra work, which is why some people prefer
subroutines or, those that use both, will use subroutines for "simple"
stuff. However the prototyping adds a lot of capabilities. WDSC has
prototyping wizards.
5 - Sometimes for certain error processing routines or whatever and you
have to send out a program message using a send program message api some
people prefer subroutines because of the stack level. But that shouldn't
equate into an "all or nothing" decision as to the use of subprocedures.
See the manual "ILE Concepts".
I forced myself to use no subroutines for a few months and only use
subprocedures. Once I developed a comfort level by
- using them several times
- using parameters and the return variable
- building my own service programs
- building a multi module program
Then I felt that I could relax and use a subroutine on occasion if
"simplicity" or the call stack issue tugged at me.


Rob Berendt