Re: the ASSET issue again

from Al Macintyre BPCS 405 CD Rel-02 AS/400 V4R3 mixed mode

Perhaps there would be value in BPCS_L working towards a collection of 
recommendations to SSA on style that they could use in their production of 
source programs that would make it easier for customers to work with their 
source code.

Once upon a time when they had modifications, the line that got changed would 
have
12345 BMR # at the side but the count got too big to fit ... perhaps they can 
be persuaded to go to an alphanumeric numbering system like A1234 B1234 with 
the first or second letter pushed up a notch when BPCS is now on a new 
version.  This would mean it would be more obvious again which lines got 
changed ... those begin end strings are real difficult to read when there are 
multiples piled on top of prior.

There are BPCS programs that are heavily saturated with literals & message 
member numbers.  It can be very time consuming for me to look up what's this, 
what's that.  I need some kind of pre-program-analysis tool that reads the 
source code & inserts a comment line in front of each of these literals etc. 
with WHAT THE HECK IS THIS, so I am not having to heavily look them up.  I 
recognize SSA uses them because of the ease of altering softcode without 
altering programs, but in 99.99% of the cases, especially error messages, 
they never do get changed.

One of the hassles I am currently working with are programs that have lots of 
numerical overflow, such that the programs are outputing BAD INFORMATION ... 
that's where MULT A.B by C.D should have a work output field of size A+C.B+D 
but doesn't because the large BPCS numbers involved are pushing up against 
the ceilings of IBM RPG and even when they are not, the way the fields are 
defined, it is non-obvious to programmer what the sizes are of the fields 
involved.  We can look this up, but it is not in plain sight.  

I do not yet have good solutions for this.  For example I need to ask 
MIDRANGE-L if there are other languages on AS/400 that are not bound by the 
30.9 limitation of RPG ... if so, perhaps when we need to MULT or DIV some 
fields that are too big to have correct RPG answers, we could call something 
written in another language than RPG, passing the values to do simple 
arithmetic.

I would like the IBM compiler to provide a warning message when there is a 
MULT or DIV or other arithmetic in which the output field is not large enough 
to guarantee a correct answer.

>  From:    Cdoe@barton-instruments.com (Chick Doe)

>  there's been a lot said in the last couple of days about SSA and AS/SET. 
> while i am not a fan of it, i can understand the logic of writing code in 
AS/
> SET that could then be compiled and run on multiple environments. yes you 
> will sacrifice performance, but if the intent was to provide improved 
> functionality at less cost, then i can understand the logic. whether this 
was 
> ever delivered or not is another question. 
>  
>  but what drives me absolutely insane is SSA's practice of commenting out 
> lines in both the AS/SET source and in the resulting RPG source. much o the 
> complexity in trying to read these damn programs is just trying to decipher 
> which lines are still executable statements and which are now comments. my 
> guess is that well over half of all statements in their programs are old 
> lines of code that are now comments. TAKE THEM OUT AND GIVE US SOURCE 
> PROGRAMS THAT CAN BE READ!
>  
>  chick doe
>  barton instrument systems

Al Macintyre  ©¿©
BPCS 405 CD Rel-02 running on AS/400 V4R3 http://www.cen-elec.com 
+---
| This is the BPCS Users Mailing List!
| To submit a new message, send your mail to BPCS-L@midrange.com.
| To subscribe to this list send email to BPCS-L-SUB@midrange.com.
| To unsubscribe from this list send email to BPCS-L-UNSUB@midrange.com.
| Questions should be directed to the list owner: dasmussen@aol.com
+---