Martin,

The REDEFINESes are not a problem in themselves. The data is identically 
represented for each of the definitions involved in the REDEFINES (the 
decimal-point is an "implied" COBOL thing).

What is a problem is how the receiving system is to know how many decimal 
places there are in that particular piece of data, because that is the only 
difference in the definitions, the number of implied decimal places. For the 
receiving system, implied is not good enough, they need to know.

Even if we were to say that the REDEFINES had two completely distinct and 
incompatible definitions, the data itself is not a problem, it fits in the same 
space (subject to short/long, where one or other of the REDEFINES items is 
padded, effectively).

For me, rather than saying "look at this, then use that field as..." it is 
better "look at this, use that field, else use another field" and better still 
"here are two fields, one will be an initial value and the other will contain 
the data, without you having to know the setting of look at this, but by all 
means cross check".

On Thursday, 4 February 2016 16:47:29 UTC, Martin Packer  wrote:
> In outline redefines could be achieved with DFSORT Symbols  and = / 
> POSITION.
> 
> Cheers, Martin
> 

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to