Todd's EDIFACT ENTREC message for General Ledger

["William J. Kammerer" <wkammerer@foresightcorp.com>]

Todd Boyle took us through a little of the EDIFACT ENTREC Accounting
Entries Message, showing "The total  possible message layouts is
Quadrillions of possible formats for each of the 6 billion people on
earth." This, when you consider all the possible arrangements of
segments, composites, elements and codes.

Todd concludes "The fundamental problem of EDI and ebXML is that you
think you can have a continuous, elastic vocabulary and grammar."

Dear Todd:

After all these months of harpooning the poor ENTREC and hours
calculating the possible instances of ENTREC, I would've thought it
possible to build a MIG (message implementation guide) of the message
which would suit your purpose.  There's even more combinations of
English (or Dutch or French or German) words to form sentences, and we
not only manage to learn to construct useful ones by the age of 4, but
also to distinguish the nonsense from the sensible.  The same should be
true of EDIFACT messages built up from its constituent components.

How many date-time constructs are useful in accounting? You tell me -
I'm no accountant.  But I'd venture a guess that only a few of the
possibilities of the DTM segment would make any sense.    Actual date
instances, Quarters, years, etc., perhaps?  Right off the bat, I bet you
can rule out anything that specifies an hour, minute or second!! How
many different types of parties are relevant to a General Ledger?  You
tell me again - I'm no accountant, I repeat.  But I bet the NAD can
identify them for the most part.

Tell you what: give us a couple examples of real ledgers, and we can
examine the ENTREC (why not the LEDGER?) and see if we can build a MIG
that accommodates your samples. I would have thought a MIG would be
available considering that someone, somewhere, had in mind what they
wanted to do before they created the UNSM (UN/EDIFACT standard message),
i.e., the raw skeleton that appears in the EDIFACT directory at
http://www.unece.org/trade/untdid/d01a/trmd/entrec_c.htm.

But in any case, examining a real-life problem like yours would be more
instructive to all of us than an imaginary one like Death registration
for ferreting out core components.  Do they really do death registration
like that described in the Core Components Discovery and Analysis Spec
anywhere in the world? Any undertakers here?  No??  At least we have an
accountant - a CPA, no less - so let's start solving your problem.

And I really don't think xCBL has solved the problem, either - if it's
really a problem at all.  xCBL has "a continuous, elastic vocabulary and
grammar" also - witness the bloated code lists that I've already brought
to your attention.

William J. Kammerer
FORESIGHT Corp.
4950 Blazer Pkwy.
Dublin, OH USA 43017-3305
+1 614 791-1600

Visit FORESIGHT Corp. at http://www.foresightcorp.com/
"accelerating time-to-trade"