Re: "If %scan(x:y)" is not valid???

Well, I "disclaimerd" anything more than one statement! <g>

Frankly, the one-statement subroutines I've seen will always only be one 
statement.  Of the
hundred or so subroutines I've seen in my career that only do a SETON LR, I've 
never seen them
changed to do more than that.

Why presuppose that the possibility exists someday that there will be more to a 
"function" than
the one or two statements that the "function" needs now?  If it does happen, 
that would be the
time to separate the function into its own subroutine / subproc.
 
Really, we could take this to the extreme and say that every line of code is a 
potential function
in and of itself that could grow.  So why not avoid the inevitable and give 
every line of code its
own subroutine / subproc?

- Dan

--- rob@dekko.com wrote:
> The problem is that the subprocedure/subroutine often grows.  And, since 
> you started it right in the first place, it grows right.
> 
> For example, I was once questioned on this list why I would do
> C/Exec SQL
> C...
> C/End-Exec
>  /free
>   // one line of code
>  /end-free
> C/Exec SQL
> C...
> C/End-Exec
> 
> But once I explained that how do you make a decision that the line has now 
> exceeded a predetermined minimum number of lines needed to justify 
> converting to free form?  Especially when people are loath to change code 
> that is working.  But if you start out right...
> 
> Rob Berendt


__________________________________________________
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day
http://shopping.yahoo.com