List Usage

["M. Lazarus" <mlazarus@xxxxxxxx>]

Simon,

At 09:54 AM 4/17/00 +1000, you wrote:

> Perhaps the ideal is a PROGRAMMING list for
> the CL, DDS, possibly REXX, OPNQRYF, SQL and 
> other general programming issues thus leaving MIDRANGE-L for the topics
> you suggest.  
> Even when those techniques are being used in an RPG application they are
> rarely related 
> to specific RPG issues.
>
> What say you all?  Should we separate the topics or alter the
> function (and name) of the 
> RPG400 list?

 I think that we should change it Midrange Programming.  Major
exceptions could be Cobol, MI and probably Java.  RPG, CL, DDS, SQL
(and to a small degree REXX) are part of the normal RPG programmer's
arsenal.  Any other midrange language volume is so low that we may
as well include it on the list.  Maybe us died-in-the-wool RPG'ers
will learn something about the "outside" world!

> I can see Midrange Dot Com taking over from
> the trade rags ... pity about the NASDAQ ... 
> the IPO for Midrange Dot Com would have been interesting
> :)

 David's probably making more off this list than most of the IPO
companies today!!  :-)


> ** A read loop should almost always be a
> DOW.  The issue being currently argued is really 
> whether a priming read should be used or a single read in an IF test
> inside the DO.  I 
> used to be in the IF camp when I first started programming but after a
> while I started 
> thinking about the construction of the loop and the additional
> unnecessary IF test and 
> changed to the priming read.  

 I never liked the extra test, so I always preferred the priming
read.  Also, after confirming w/ the compiler developers that the
"additional" READ opcode doesn't add much to the program size,
and the performance s/b better anyway w/o the extra test, that clinched
it.


> ITER and LEAVE (and the appalling LEAVESR)
> fall in to the 
> same category and I've said enough about them in the
> past.

 I don't see how LEAVESR is worse than a LEAVE.  I think it's
better, since the scope is usually smaller and there is a definite exit
point.  It's clear and even legitimate in structured
programming.  A "we're outta here" statement to a single
pre-defined exit point is fine, IMHO.
 
 -mark