Re: Adding columns to existing Infor files.

For every standard we found out about, there were exceptions.

Not all BPCS Business rules are in one place, but fortunately not a lot of
different places.
Some files have multiple members, and sometimes a hassle to manage ones we
are done with.

We were told that the 200 series reports were simple reports, while the 400
were compound.
Well we found a lot of 200 series, which I would have called compound, where
the first stage determines selection criteria and report sequence, then
creates a work file containing the needed data, sequenced parallel to the
needs of the final report, which was processed from that physical, not any
logical.

I recall the calendar file had 1 record per year, with string of W-weekends
& other codes, but invariably got converted into a different format file
before being used for any function other than defining the calendar.

Certain letters of the alphabet designated files, menus, etc. for particular
applications. But there were exceptions to the rule.

Most of the logical are named using the name of the physical, the letter L,
then a couple digits.
Watch out for the occasional letter J instead of L. That is a logical which
is joining two files.
That is one way to handle the addition of some fields, instead of your IIM2
example.

IBM OS had GO CMDREF which got me to a menu where I could run a
cross-reference of which programs, in theory, accessed which files, but it
listed them by logical, not the raw physical, and sometimes the CL was
making substitutions to what was in the source code. So I knew this was not
the whole story, but then I put that in an *OUTFILE accessed by query, where
I could designate the 3 letter physical file, which included all logical and
joins, named after the physical, to get a list of pretty much all the
programs that did certain actions with any variant of that file.

There is a complication when there is a need to reorganize any of the 2
files. (In theory more than 2 files can be joined with a single J, but I
found IBM rules for sequencing a J file based on controls in only one of the
physicals, to limit this technique.)

We have to know ALL the places where BPCS might want to do updates to the
physical, which requires action with the logical. We were rather
disappointed with the XRF application for BPCS 405 CD, which included some
security issues.

3rd party XRF would not be able to handle how BPCS managed the L-messages,
where the software code did not have actual text, but substituted L-messages
at run time, in case of multi-linqual environment. I had my suspicions that
architecture was not being used exclusively for that purpose.


Alister Wm Macintyre (Al Mac)