Why will EDI keep on existing along with XML?

It has now become a platitude in the computing press to always position EDI against XML as B2B data exchange technologies.

There is no doubt that EDI is a widely spread protocol but it still presents a latent defect since it is a proprietary format and, as a result, exclusively depends on intercompany agreements.

This is why it would be very difficult to adapt EDI to the new requirements that are made in the B2B data exchange information systems, which are more and more part of the Internet core business (integration to the customer and provider marketplaces, Web sites, development of Web Services structure, use of services in the ASP mode).

What's more, due to its nature, EDI would only be used in big enterprises and would give them the possibility to keep those of their providers who are involved in this type of IT and who could hardly do without it…

On the contrary, XML would be viewed as a universal data exchange technology and would carry the hopes of a coming new age for IT intercompany relations, deprived of any proprietary aspect.

Just as the other Internet technologies, XML development no longer depends on the enterprises themselves but on the Internet developer community, which should facilitate data exchange in the BtoB sector.

What's more, it would be a way for XML, whose specifications are still wide open, to reduce the integration costs of the intercompany information stream, giving the small and midsize organisations the opportunity to connect themselves to the information systems of the big principals.

In this market binary vision that imposed itself massively in the last two years in computing reviews, XML should rapidly replace EDI as the reference technology used between companies to exchange information. As for EDI, it should soon disappear into nothingness as far as IT technologies are concerned.

It must be said that software vendors largely promoted such angle and only banked on XML! Considering how important their advertising and marketing budgets are, it is somehow easier to understand the influence they might have had upon the specialized press and opinion leaders to promote their own vision in the battle EDI versus XML.

But reality is stubborn. And today we are forced to the conclusion that developments in the field of intercompany exchange do not always correspond to what was expected to happen.

The first thing that prevents XML from quickly displacing EDI is the fact that so many enterprises use EDI in some branches of industry.

This is how more than 90% of the actors involved in the mass marketing and the car industry sectors have adopted a standard such as EDIFACT, dialect that took them 20 years to achieve.
We can easily imagine how much work was involved in the process, not only on the intellectual level, but also on the computing and organization level.

Nowadays, EDI has become a recognized standard in these branches of industry, well mastered in both its set up and results.

Enterprises are not trying to please software producers, even the XML fans, but are only trying to secure their daily transactions using proven technology.

What's more, as most of EDI investments have now been paid off, the cost has been much reduced and only involves developments which are referred to as technical maintenance. Besides, enterprises presently have many people working in their IT department who master such environment perfectly, which is far from being the case for the new developments which are based on XML technology.

But what really seems to favour a long-lasting coexistence of the EDI/XML technologies is the conception mode used between those two data exchange formats.

If we look at it in an objective way, we notice that their maturing process is nearly identical!

Contrarily to the image that might have been conveyed, one must know that EDI is not a technology that was established through the sole efforts of its editors who would have since then become the only masters of its evolution.

During the past twenty years, EDI has been widely implemented and is now highly integrated into core business processes.

In the most compact branches of industry, mainly the car industry where so few actors are involved, the said actors took hold of the matter and managed to implement some common EDI procedures.

What about XML? The same procedure was used!

And this is easy to understand: each branch of industry needs to define a dictionary containing the specific words they have in common, and the only way to do so is by putting together the main enterprises of a given sector that will manage to agree on common definitions that will later impose themselves on the rest of the actors.

And this is an endless process since needs keep on evolving.

And yet, only the biggest enterprises are able to do this standardization process since they are the ones that most urgently need it but also the only ones who can afford such heavy investments.

With XML, they are asked to start over and do again what took them 20 years to install. It is then easier to understand why they feel somehow reluctant to migrate from one technology to the next.

The first set-ups that took place on some marketplaces well prove it.

In sectors where EDI prevails (car industry, mass marketing), we notice that only the biggest marketplaces, which are promoted by unions of actors coming from the same sector, can stand up to independent marketplaces.

The nature of the exchanged information (product order, reporting, invitation to tenders) and their level of integration to the production processes (cf: just in time production) have become too strategic for the concerned enterprises to let a third party manage these productions.

Such comment does not apply to the sectors in which the demand is so widely broken up that no actor or even group of actors have any interest in federating wishes to define a common language at their own costs (packaging or transportation sector, for instance).

In those sectors, it might be possible for one or two independent actors to impose their own marketplaces.

As for the editors, we can also observe that we are far from having a XML that would be considered as a universal technology "spontaneously" developed.

From UDDI to Biztalk (Microsoft), via OASIS (supported by IBM and Sun) and ebXML, there have been many attempts to promote a professional XML, attempts that were always prompted by a group of publishers and that were not carried out in the Open Source way as was the case of the other Web technologies (HTML, PHP, etc…).

At the same time, as XML is a very simple technology (this is a strong argument, any member of staff can understand it and not only members of the IT department), it allows small and midsize organizations to get into connection without having to resort to a heavier technology, such as EDI.

Besides, XML development prospects remain high, especially the unionisation of web content…. but maybe not on the points where EDI proves very satisfying!

As you can see, XML revolution should take place in a progressive way in the sectors where EDI is most widely spread.