Michael Mattias said:
>
> I'd think "validation against an IG" rather than "against the the X12
> standard " might entail a couple of different "levels" of compilance...

Certainly there are other things that you can do which will greatly enhance
the value of the syntax checking.  Many of these, of course, are pointed
to by the X12 Technical Report on "Compliance with X12" which lays out
what IGs may or may not do while remaining compliant with X12, and similarly
data messages within both the X12 Standard and the IG.  You can get that
Technical Report from DISA.

> 1. Required/optional/forbidden segments at various levels of the document.
> 'Required'  could include "minimim/maximum required (per loop ID or document
> position)".

The proposed 997 changes had new codes for all of this, plus of course
the same codes for elements and repeating elements.

> 2. For required/optional/not permitted segments, " business data" conveyed.
> For example, an IG might read that (using a HIPAA-like example) two NM1
> segments are required: One to identify the patient, one to identify the
> facility.
>
> You might see the same thing in an 850 PO: two DTM segments required, one
> with "ship not before," one "ship not after."
>
> 3. Code values accepted for item type ID; may vary with the location of an
> element within the document.

These changes were also in the proposed 997 changes, but are now in TS 824,
because it allows context to be described - that is an error report can
not only describe the error itself and its exact position in the EDI
stream, but also the position of a preceeding piece of EDI that made
the current data invalid.  This is like saying:

   because you used code ABC at DEF.04, this code GHI at JKL.07 is invalid

> 4. Then the trickest one, because there is no way any "off the shelf
> commercial translator" to handle it: when the IG says, "Must be a valid
> 800-lb Gorilla, Inc. Purchase Order Number"

That depends on exactly the purchase order looks like.  Some off the shelf
translators (Mediator by Syscomm Strategies, for example) can do this sort
of thing because they allow significant integration with the application
environment - so they have code that can actually go and look things up
in industry standard files (C-ISAM, ODBC, etc).

But certainly most translators ought to be capable of simple edit checks
against a COBOL-like mask:  99-999999-X or the equivalent Unix regular
expression: [0-9][0-9]-[0-9][0-9][0-9][0-9][0-9][0-9]-X.

I can say that our VAN software and translator products support this kind
of operation, powered as they are by IMPDEF, the international EDI Standard
for describing IGs in a machine-processible form.

Jonathan
------------------------------------------------------------------------------
Jonathan Allen             | [EMAIL PROTECTED] | Voice: 01404-823670
Barum Computer Consultants |                             | Fax:   01404-823671
------------------------------------------------------------------------------

=======================================================================
To contact the list owner:  mailto:[EMAIL PROTECTED]
Archives at http://www.mail-archive.com/edi-l%40listserv.ucop.edu/

Reply via email to