[EDI-L Mailing List Archive Home] [Message List] [Reply To This Message]

Re: From webMethods to GIS: A Painful Odyssey

Re: [EDI-L] From webMethods to GIS: A Painful Odyssey

What you're describing is used often in mapping systems when the inputs
and outputs
are diverse and they share a common data model. For example, in a
translation model
in a service bureau, the input is X12 and the output is the customers
format (idoc, edi,
xml ...whatever). The X12 is mapped to a common format in the middle
(the canonical)
and then remapped in the second pass to the custom output format. This
model really
works when it scales, because you only have to add plug-ins on either
end rather than a
new map for each new ("finished") transaction. The reason that
webMethods uses a
canonical model is because webMethods is nothing but a transport - it
takes something
from one place and puts it into another. The canonical really works
well as xml, because
it is easy to maintain and modify.

In an end user environment, the canonical model is a waste of time and
effort. Why?
Because one of the ends - the one on your end, in this case the idoc -
doesn't change.
You'll use a common format model for the idoc, and have multiple maps to
read X12 to
produce the idocs. Or to read the idocs and product the customer's
edi. The TN model
of creating maps with plug-in modules is unique, but a map is no longer
a stand-alone
program, with it's own unique stability and identity. Change a TN map
and get careless,
and you'll bring your whole system down. The only knock against gentran
is the one file
in, one file out mentality. With other translators, you can read and
write to multiple files
simultaneously. What you want to try and do is minimize the number of
maps, especially
on the inbound side. If you only have one inbound map, then you really
don't need a
canonical. What you also want to accomplish is version neutral data
structures. Idocs
are just that - version neutral.

If you really want to replicate the WM experience, you'll need to design
(or steal) a common
interface format, and run either imbedded maps or two translation
passes. Considering how
baldly MW tracks and reports translation passes today, why would you
want to introduce
more failure points?


Ron Paquin wrote:

>Third week on the new job and though I would have denied the
>possibility, I find myself missing webMethods.
>
>Specifically, in EDI in webMethods, we used an approach involving so-
>called canonical documents (also known as Common Business Document or
>Universal Data Model).
>
>For example, and inbound X12 4010 850 was first translated into a
>Purchase Order Canonical, and then that Purchase Order Canonical was
>translated into an SAP 31i IDOC.
>
>Why? A number of reasons. From the Purchase Order Canonical, we could
>easily translate into any back end system. For example, Kroger sent
>us POs for our core business, but also for an acquisition that had
>not been integrated into SAP. So the X12 850, for either destination,
>was translated into the PO Canonical, and from there the core stuff
>was translated into IDOCs, and the other stuff translated into XML
>and from there, via an XLST, into a PDF to be emailed to the
>acquisition folks. Now, once the acquisition is integrated into SAP,
>our only change will be to use the same path to the IDOC that we do
>for the core stuff--literally, a five minute change.
>
>Another example. We initially implemented Wal-Mart at 4030, but
>they've recently mandated upgrading in 5010. In any X12-to-IDOC
>translation, this would require recreating all of their maps in the
>new version--several weeks' work. In wM using this approach, however,
>it was simply a matter of changing the schema in the existing map
>from 4030 to 5010 and repointing the mappings--a single afternoon's
>work.
>
>I'm probably preaching to the choir, if anyone reading this is
>familiar with webMethods and their GEAR documentation. There's a
>White Paper called "Canonical Strategy" that you're probably familiar
>with.
>
>At any rate, what I'm looking for is some blessed person who has gone
>from webMethods TO GIS, and who can point me in the direction of how
>the heck I replicate this functionality in GIS.
>
>Anyone?
>
>Regards,
>
>Ron Paquin
>Tenneco Inc.
>
>

--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Brian Lehrhoff, EA ( Messaging Consultant 201-913-4506

Buy 2006 Quickbooks for 15% off at http://www.brianlehrhoff.com
Circular 230 Notice is located at http://www.brianlehrhoff.com/230notice.html
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%





Fri Mar 10, 2006 6:05 am


Brian Lehrhoff <blehrhof@...>
http://profiles.yahoo.com/blehrhoff
Offline


EDI to XML Mapping for EDIFACT/X12 Convert EDIFACT/X12 Schemas to XML Schema Legacy Data Conversion Tools Access Relational Data as XML Visual XSLT and XQuery Mapping Tools Simplify EDI Data Integration with Stylus Studio XML Enterprise Suite - Free Download!
Subscribe in XML format
RSS 2.0
Atom 0.3
Site Map | Privacy Policy | Terms of Use | Trademarks
Free Stylus Studio XML Training:
W3C Member
Stylus Studio® and DataDirect XQuery™are products from DataDirect Technologies, is a registered trademark of Progress Software Corporation, in the U.S. and other countries. © 2004-2007 All Rights Reserved.