|Date: ||Wed, 28 Feb 1996 10:56:54 EST|
|Reply-To: ||Gerry <STATMAN@PACEVM.DAC.PACE.EDU>|
|Sender: ||"SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>|
|From: ||Gerry <STATMAN@PACEVM.DAC.PACE.EDU>|
|Organization: ||Dept of Academic Computing, Pace University|
|Subject: ||Overlapping SCL Sections|
Roger Bolton writes ...
>Stacking Labeled sections has always seemed to be controversial.
>What I find dangerous,is when the following is done
> A = ;
> B= A ;
> return ;
I second Roger's position. Although I only have minor experience with
SCL, I've run into code in other languages that allows a similar type
of technique to be used.
Both PL/I and FORTRAN have the ENTRY statement, which allows a second-
ary entry point into a subroutine, which is essentially what is being
presented in the code above.
Its a bad idea in PL/I and FORTRAN, and its a bad idea in SAS/SCL,
particularly in view of the rather vigorous discussion of spaghetti
code vs a structure analysis/programming approach that took place
a couple of weeks ago.
Use of this type of coding results in problems that are not unlike
those encountered by excessive use of GOTOs, as well as making errors
more difficult to diagnose than they normally would be if this
technique is avoided.
Just my two cents on this subject !
Gerard T. Pauline
Programmer / Analyst
Dept of Academic Computing
Pace University, NYC Campus
Statman @ Pacevm.Dac.Pace.Edu
Statman @ Pacevm