William,

I fear my point (probably overly simplistic) has been obscured in the 
discussion.  (On the other hand, maybe I just don't get it...)

My overriding point is that examples that illustrate a straight one-for-one EDI 
to XML conversion mean exceedingly little since the primary bugaboos of EDI are 
really not solved.  Though I don't "laugh away the parsing problem as trivial," 
I respectfully regard the parsing as the trivial part of the equation.  Perhaps 
there is a terminology problem.  The *parsing* of EDI is a strictly mechanical 
and predictable operation that frankly doesn't require an expensive translator 
(in fact, the syntax of EDI is arguably a little simpler than XML).  To me, 
*parsing* is just getting at the data.  The magic part is the *translation* 
which figures out what data (once it's been isolated by parsing) goes where in 
the application.

Far be it for me to suggest XML is a bad idea.  It definitely can replace much 
of what EDI does, but it's not a drop in replacement, it's (dare I use the 
clich�?) a paradigm shift (I dare, I dare!).  Your examples illustrate that 
very thing.

Respectfully submitted,
Bill Chessman
Inovis(tm)

-----Original Message-----
From: William J. Kammerer [mailto:[EMAIL PROTECTED] 
Sent: Monday, January 31, 2005 3:59 PM
To: EDI-L Mailing List
Subject: Re: [EDI-L] XML Hierarchy Question


Bill, I myself have asked the same question: "what's the point?"  The
intent of mechanical EDI-XML conversion isn't to "bulk" up the data for
transmission over the network just for the heck of it.  The same old EDI
is sent between companies.  But once it's received, it can the be "blown
up" into XML using, say, Bryce Nielsen's xmlLinguist.   Yes, "one would
still be faced with all of the old problems that have dogged
EDI...semantic content remains", and on and on.  But at least the
*parsing* problem goes away - or at the minimum it can be done without a
$50,000 translator (where the only support you can find is by bugging
folks on the EDI-L mailing list).  That's the theory, at least.

Don't laugh away the parsing problem as trivial.  Perhaps upwards of 80%
of inter-company bulk data transfer is non-EDI (in the X12 and EDIFACT
sense).  Most of it is flat files, comma-delimited records, and stuff
like that.  None of that stuff is going to be turned into classic EDI,
even if X12 and EDIFACT got busy devising messages for each and every
possibility.

But it's conceivable that if we e-business folks devised a workable
framework for core components implemented with XML and schema, then
perhaps the non-e-business folks would see the merit of following in our
footsteps, and take all their data (anything from oil drilling geologic
data to toxicological reporting) and do the same. With almost 100%
certainty, I can state that these other data sets will **never** be
encoded in classic EDI.  If we couldn't convince them before to use X12
or EDIFACT, we're certainly not going to do so now (since everyone has
heard of XML).

Let me tell you a short story.  I had some chunks of data once, encoded
in ASN.1 BER  and described by an ASN.1 specification (much like an XSD
schema).  When I told the programmer at the other end what to expect, he
was not in the least perplexed.  He vaguely knew of ASN.1 and even heard
of the freeware SNACC routines for parsing the data.  That's the last I
heard from him regarding the syntax, and I confirmed that he was able to
read the data relatively effortlessly.  Mind you, ASN.1 is not at all
common, but there's a heck of a lot better chance that the random
programmer on the street is going to at least have heard about ASN.1,
more so than the X12 or EDIFACT syntaxes.  This was complex recursive
data which would not have been suitable in some comma-delimited format,
by the way.  I'm not advocating ASN.1, as anything you can do in ASN.1,
you can do in XML better.

By the way, with ebXML core components, you're not limited to
implementing the physical data in XML - even ASN.1 encodings can be
used.  The same core component models used to build the XSD schema were
also used to generate the ASN.1 specification!  See Appendix F
(Informative): ASN.1 Specification, at
http://docs.oasis-open.org/ubl/cd-UBL-1.0/.  Amazing, huh?  Try to do
that with X12 or UN/EDIFACT.

William J. Kammerer
Novannet
Columbus, OH 43221-3859 . USA
+1 (614) 487-0320

----- Original Message ----- 
From: "Bill Chessman" <[EMAIL PROTECTED]>
To: "EDI-L Mailing List" <[email protected]>
Sent: Monday, 31 January, 2005 12:44 PM
Subject: RE: [EDI-L] XML Hierarchy Question



While the points made about XML below (ie., one XML root document per
file, the hierarchy questions, etc.) are quite correct, it always
troubles me to see examples that at least suggest a one-to-one mapping
between an EDI document or interchange and XML.  If that's all that a
given product can do (reformat the EDI into pretty "human readable"
bits), I must ask, yet again, "what's the point?"  Assuming that someone
actually did such a transformation one would still be faced with all of
the old problems that have dogged EDI...semantic content remains
unclear, coded fields remain coded and their respective meanings are
those of the Trading Partner (not necessarily in alignment with the
standards themselves), to name just two.  On top of that, you expand the
size of the data (I expect disk drive manufacturers have no problems
with that).  The task of getting the data from the documents to the
application remain the same (some sort of map must be created).

Maybe I haven't been paying attention, but I thought the point of XML
was to foster interactivity with shorter messages that work much more
like a dialog.  (I think that's one of the goals of ebXML and
WebServices...in fact, ebXML and RosettaNet both use MIME wrappers
around their XML data, don't they?)  Although there is also Interactive
EDI, it still hasn't taken the world by storm, so the vast majority of
EDI remains batch oriented.

Sure, I understand that SAX is very easily obtained, inexpensive or
free, and pretty easy to use (I have used Java versions of SAX on
numerous occasions).  Sure, I understand that using XSLT you can
reformat the XML data in an assortment of ways.  From a user perspective
it sure is nice to see all that icky data presented in a pretty HTML
format in your browser, but the point of the EDI (which XML seeks
to--and may still--replace) is to transfer data from application to
application.  Sure, I also understand that you can use XSLT to reformat
the data so that it can be fed into your application (back in the olden
days, we called that process "mapping").

Okay, I guess I've been ranting a bit (gee, I don't think I've ever done
that before! 8-) ).  In summary, if EDI is so objectionable that we need
to move toward XML, we need to think more about the process than the
format of the data.

Best regards,
Bill Chessman
Inovis(tm)



.  
Please use the following Message Identifiers as your subject prefix: <SALES>, 
<JOBS>, <LIST>, <TECH>, <MISC>, <EVENT>, <OFF-TOPIC>
Access the list online at:  http://groups.yahoo.com/group/EDI-L
 
Yahoo! Groups Links



 




.  
Please use the following Message Identifiers as your subject prefix: <SALES>, 
<JOBS>, <LIST>, <TECH>, <MISC>, <EVENT>, <OFF-TOPIC>
Access the list online at:  http://groups.yahoo.com/group/EDI-L
 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/EDI-L/

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 



Reply via email to