The future of EDI

[longhair]

all,
being that this is a 'general discussion' group...
what are your thoughts about the future of 'traditional' EDI (x-12, edifact).
since i started in the arena i've seen not only numerous spec / version changes but also communication protocol changes.
these are all pretty minor when you consider that there are some who want to revamp the entire concept (read xml).
thoughts? preferences?
regards,
longhair

 

[JuJutsu]

Personal opinion: XML for EDI is a waste of RAM and bandwidth. EDI as it stands will be around long after anyone reading this will have retired.


JuJutsu - Jeff S.
Support Analyst

 

[longhair]

JuJutsu,
that is one of my issues with xml.
regards,
longhair

 

[Dimandja]

I am new to EDI, but I am beginning to have some thoughts about it.

1. EDI is essentially an old technology. Read "older than XML".

2. Using XML does not mean manually coding all those tags. XML technology has evolved to the point where an EDI transaction could simply look like a function call, or a method invocation.

To me XML would be a vast improvement to the current positional text being ferried back and forth. Not to mention the tedious mapping.

 

[longhair]

Dimandja,
i do not think that is true, if you do not include all of the xml tags you will have to map, most likely for every trading partner. one of the things that has always bothered me, about edi and business is the creation of a spec that no one completely adhears to. they all say, use this spec but change this aspect of it, because we are different. i see it as just a way to get your partner to correct the tp's internal issues rather than them correcting it themselves (read passing the cost).
from what i've seen of those that want to use xml, they include all of the tags. this, i believe, is what JuJutsu was refering to.
regards,
longhair

 

[JuJutsu]

EDI (example) GS...,,Smith,John,J,,...GE Data is known by two byte initiator and position in the data stream with one byte delimiters.

XML (example) </up to 3 megs of empty tags required to get to your data><LAST_NAME>Smith</LAST_NAME><FIRST_NAME>John</FIRST_NAME><MIDDLE_INITIAL>J</MIDDLE_INITIAL></up to 3 more megs of empty tags>

;^)

JuJutsu - Jeff S.
Support Analyst

 

[Excalibur1701]

I've heard the "EDI is dead" story 7 years ago. Since then, I have doubled the amount of trading partners and significantly increased our transaction flow. The EDI direction is dictated by the big companies. For us, that would be Cummins, Mack Trucks and Volvo. Until they change their practices, EDI will be around for awhile.

 

[golancy]

Hi Excalibur,

What you said was true.Even I work for Cummins EDI. Until these companies would ah\dhere to EDI it will stay.

Cheer's
Arvind
arvindkumar.raju@gmail.com
India

 

[damber]

I work for a large global corporation that has a customer portfolio that looks like a global business directory - we work with most major non-financial company across the world (FTSE100/FORTUNE500). So what have I seen in terms of Integration requirements and trends in regards to EDI ?

Well, EDI is far from dead. XML is far, far from standard (at least in the general implementation of the concept), and as longhair noted - the problem is not the 'standard' that is used, it is the underlying business process - businesses just want to be different for the sake of it, in some feint delusion that it is in fact a differentiator.. when they know that it is in fact just a cost factor they are not willing to address.

EDI is a much more efficient data format than XML - XML is for people, EDI is for machines. A machine has no preference whether the element is encapsulated by a 'pretty' name tag - but it does make life easier for humans.

XML is also a very open standard... a format for making formats... just because someone uses 'XML' doesn't mean that your XML interface is compatible... In fact, I would even go as far as saying that just because someone uses EDIFACT/X12/TRADACOM, doesn't mean that their messages can be understood by an existing interface into your application. Sure, it will interpret the syntax and structure fine, but it will not necessarily understand the content fully. Simple example of this is the phrase 'Order Number'.. this can have many connatations (Customer Order Number, Client Order Number, Sales Order Number, Purchase Order Number, Original Order Reference, Picking Request Number, Shipment Order Number... etc... etc).. People interpret these things differently. Though standards do help to rationalise the options a little.

Rationalisation of the myriad standards is essential. Traditional EDI is already split into several camps.. US: X.12, Old UK based format: TRADACOM, and the most international version (but mainly Europe): EDIFACT There are others but they're not worth mentioning ) So that is before you start to compare XML based standards such as ebXML, OAGIS BODs and RosettaNet.

The thing about most of these standards is that they are data formatting standards.. the sender encodes the data in a way that the receiver can read it and manipulate it.

The only standard that really addresses this flaw and is a true process based standard, which happens to include a data format to support it is RosettaNet (

http://www.rosettanet.org).

This emerged within the Technology industry, but has started to move into other sectors too - our most recent non-tech sector request came from the worlds largest pharmaceutical company.

If you want to know what 'big-business' is working towards, then RosettaNet is it - for the technology sector at least.. it has almost already happened - for others, this may or may not emerge, though I would say it will start to make serious advances in all areas, as it also supplies a process framework, which (although requires additional investment) helps businesses work together more efficiently. Have a look at some of the companies that are involved in it, you may notice some of the smaller ones, like Microsoft, IBM, Intel, Nokia, Cisco, Sun, HP, BT etc ;o)

It uses an XML based data format, and transports over internet protocols (e.g. HTTP(s)) and includes non-repudiation and security techniques, e.g. digital signatures, encryption etc as part of the RNIF framework (as standard)

So, the various bespoke-XML standards that people 'make-up' unlikely to ever succeed. The various XML standards other than RosettaNet are still struggling to make a useful difference (though BODs are noteworthy). The older EDI standards are message centric and limited, but are the most popular because people 'know' them better than the newer standards. So -at a corporate level EDI is current king, RosettaNet is set to charge the throne in the next 5 years, though other formats are still involved - especially in various industry verticals, take the German Automotive industry - VDA is pretty much the one and only standard! (unfortunately). Smaller businesses, who are new to integration, are more likely to opt into simple XML formats, as this is more commonly (and easily) understood and implemented (due to the widespread use of web technologies, including XML ).

Another topic is webservices - a young and emerging technology / concept - but far from where most businesses are. The Service Oriented Architecture may be the buzz of the moment, but the reality is a way off yet - check back in 3-5 years.

Personally I think there are more important factors to be considered than the format of the data - getting the business processes simplified, rationalised and standardised is the first and foremost. Add to that the differentiation of 'machine' information and 'human' information - machines don't need lots of extra tag information and whitespace.

Another favourite topic of mine for improvement is common information language and representation (also known as a semantic ontology) - what is an order number to you ?

And finally, people need to learn how to use standards correctly - so often you will see an implementation where an element or segment has been abused due to lack of understanding of it's purpose.

But who knows - maybe we'll go back to using carrier pigeons and the like...

Hope these ramblings help you in whatever quest you have embarked upon to make you ask such a question..!

A smile is worth a thousand kind words. So smile, it's easy!

 

[longhair]

damber,
nicely written.
regards,
longhair