I've returned from a recent set of roundtable discussions in which several retail clients heatedly discussed the migration of existing EDI (electronic data interchange) systems to XML-based alternatives. After intense discussion, I found that the number of proponents of the idea was roughly equal to the number of its detractors.
The same was true of feedback I got from my recent article on EDI standards. Whereas many readers wrote to tell me how in love they are with EDI systems and how reluctant they are to trade them in, just as many wanted to know why companies that use these "prehistoric systems" don't wake up to the future of XML. Why continue putting Band-Aids on a technology that never captured a very large market share to begin with?
Mike Barton, EDI coordinator at retail clothier Brooks Brothers Inc. in New York, indicated that his reason for staying the course with EDI and VANs (value-added networks) was rooted in the level of security it provided. The same is true of many companies.
XML may well be the future of business-to-business computing and online trading, but one must take notice of the irony of asking a company to throw away decades of EDI development on an XML alternative that is not winning acceptance from any standards bodies. XML has more than 100 fragmented variations available.
In consideration of the original investment cost, most companies with large EDI systems are, rightfully, in a holding pattern to fully investigate the business benefits and market shakeout of XML before signing off on an XML implementation.
Furthermore, although there certainly are implementations of interactive EDI and batch XML, these are by no means traditional treatments. XML transactions are typically real-time, ondemand collaborations, whereas EDI typically accumulates transactions before moving a batch file to its destination. For that reason, most company's EDI implementations are transporting large volumes of data from point to point.
If the data being transported contained information that affected the bottom line of my company, I would want to know that some service-level guarantee was going to ensure my data's arrival before I started prying open links in my supply chain. You're not going to get that type of throughput guarantee from an ISP -- not over this Internet.
As with EDI, XML standards are taking multiple routes. Although the base specification is well managed by the World Wide Web Consortium (W3C), variants are churning out quickly across different industries.
New specifications are constantly under way, such as ebXML (e-business XML) from the United Nations Center for Trade Facilitation and Electronic Business and the Organization for the Advancement of Structured Information Standards. The groups aim to provide a framework for a global marketplace based on a common set of XML semantics. Although ebXML has only officially generated a requirements specification, the effort is demonstrating impressive momentum, and the groups recently announced broadening their messaging specification to include SOAP (Simple Object Access Protocol).
It's likely that we all know where this road will inevitably lead. But there are challenges that XML must face before many organizations will consider adopting it. Although XML will continue to gain acceptance as the great equalizer in facilitating b-to-b collaboration, for many companies, letting XML cool off a bit may prove a prudent strategy.
But before we trounce EDI's capability of holding its own in an Internet economy, let's appreciate the benefit of knowing the technology we have rather than betting the farm on one that's not yet here.
For the record. In my Feb. 19 column, reference was made to the group responsible for developing the EDIINT (EDI over Internet) AS2 standard, namely Dale Moberg, Dick Brooks, and Rik Drummond. Let me clarify that Brooks is CTO of Group 8760 in Birmingham, Ala., a company devoted to helping companies with "a sunk investment in legacy EDI value-added networks." Drummond is CEO of the Fort Worth, Texas-based Drummond Group, which provides b-to-b strategy and XML/EDI related support.
James R. Borck is managing analyst at the InfoWorld Test Center. You can send him email at firstname.lastname@example.org.