XML's One-Two Punch

by Bill Trippe

People often ask me if XML is over-hyped, and I always answer the question with my favorite industry joke: What is the difference between a car salesman and a computer salesman? The car salesman knows when he is lying.

In other words, yes, it is over-hyped, often by people who don't understand XML as well as they should. The result is that XML is often given as the answer to every question. As one industry article recently observed, XML is "the software industry's latest solution to world hunger."

Are those banking on software infrastructure based on XML heading for disillusionment? There do seem to be some warning signs on the horizon.

Hype itself is a problem. It doesn't take long for smart people to tune out hype and for skepticism to grow when early adopters don't reap immediate (and huge) benefits. Indeed, there is so much noise around XML that its real benefits — such as data normalization, validation and interchange — end up seeming bland in the face of the hype.

Another problem is the proliferation of XML vocabularies — the many initiatives and schemas that have been developed to express data in XML format. These range from EDI-style business documents to scientific data to supporting data and message formats for software development itself. At this point, it would take a team of analysts to develop a comprehensive and accurate catalog of all the XML initiatives out there.

The threat of so many competing XML initiatives is that interoperability will become impossible. If n groups devise n ways of expressing, say, genomic data in XML, how will company A and company B end up successfully communicating with company C? The result, some fear, is a "Balkanization" of the XML community, where competing factions end up speaking different languages.

I really don't worry too much about either Balkanization or hype. Why? Because the good news about XML decisively outweighs the bad. In these days of renewed interest in profitability and return on investment, XML-based development is clearly the way to go.

To begin with, XML-based development is your best chance to preserve existing systems while bringing business processes to the Web. XML is the lynchpin in Web services — the best means by which legacy systems can communicate with your users via the Web. The buzz around Web services is significant precisely because it represents the best of both worlds; you can leverage the existing systems you choose to while bringing whole new groups of users to your business via the Web.

Take, for example, the U.S. Air Force, which is using XML to drive the electronic distribution of maintenance and flight manuals for E3 AWACS aircraft. The documentation had been maintained for decades on proprietary systems that produced paper output. When the Department of Defense mandated paperless delivery, Crystal City, VA-based military contractor Veridian developed an XML delivery platform using Ixiasoft's TextML Server. The legacy data is still maintained on the proprietary systems, but is also converted to XML for search, retrieval and distribution. The retrieved data is converted to HTML on the fly and displayed on a customized browser. The result is a low-cost system that lets the Air Force preserve its legacy system while delivering up-to-date paperless manuals.

Secondly, XML is about intelligent use of local data so it can work with other internal and external systems. The fallacy of the Balkanization argument begins with the fact that your organization likely does speak some of its own language. There are valid business reasons why your version of a purchase order may be different from the external versions you share with partners. You may not want to share all internal data, or at least not share it in the precise form you maintain.

The real beauty of XML is the ease with which local data can be merged with or converted to external data. You end up with the local data you need, and you get a low-cost, flexible means to communicate with external parties.

An engineering firm I work with recently decided to outsource print-on-demand reporting. The reports themselves are maintained as XML source data on one system; pricing and shipping are maintained in an Oracle database on another system. When the engineering firm needs to print and ship a report, it extracts and formats the report from one system and draws the raw pricing and shipping data, converting it to XML format, from the other system. The report formatting and XML data are transmitted to the outsourced printer over the Internet. The report and invoice are packed and shipped the same day. The engineering firm has outsourced a vital function at a cost savings; the printer efficiently serves its customer.

Together, the two benefits of preserving existing systems and improving data interchange are a significant one-two punch for XML. In fact, these may be the most important things that technology projects can do for organizations today. Many legacy systems do work, yet they lack contemporary user and programming interfaces. XML-based development will give these systems exactly what they require.

Bill Trippe (btrippe@nmpub.com) is president of New Millennium Publishing (www.nmpub.com), Boston, a consultancy specializing in electronic publishing, content management, SGML and XML.