Re: CHAIN + DS as result field

[Adam Glauser <adamglauser@xxxxxxxxx>]

Brian Parkins wrote:
> > > For the record I also see no value in using %kds.
>
> Just a thought.  What if the key were to change at some point in the 
> future?  Only one change to make, (on the D-spaces, where everything else 
> is defined) vs. many in-line changes on the C-specs.

On the other hand, if you are reading the same file with the same set of 
keys in that many different places in the program, perhaps the program's 
design needs to be reconsidered.  Not only that, but if the key is 
changing there's a good chance the business logic surrounding those 
reads is going to be changing as well.


> If the following were also used:
>
>      D KeyDS         E Ds                  ExtName(FileX:*Key)
>
>         Chain %KDs(KeyDs) FileX;
>
> Then code changes might be avoided altogether - just a recompile 
required.

I see the attraction of this, but don't forget that you'll also need an 
"automatic" way to fill the fields in the KeyDS.  The only thing I can 
think of off the top of my head is eval-corr (see James Perkins' earlier 
post http://archive.midrange.com/rpg400-l/200902/msg00212.html).