Speaking the Same Language

he builders of the ancient Tower of Babel would have felt the pain of today's federal technology manager.

According to Genesis, God confounded the tower's architects by forcing them to speak different languages. Unable to understand each other, the builders quit their project for the grand edifice.

The story of the tower haunts federal agencies today. Over the years, they have built massive stores of electronic data in such unique ways that one agency's information often cannot be shared, used or even understood by other organizations. The Sept. 11 terrorist attacks showed the public that many law enforcement and intelligence agencies weren't sharing leads about suspected terrorist operations. In some cases, laws kept them from doing so, but another big hurdle was technology. Just like the builders at Babel, many of those agencies wouldn't have understood much of each other's data, no matter how much of it they swapped.

Now, some federal technology gurus hope they've found agencies' modern-day Rosetta Stone, the ancient Egyptian tablet that unlocked the meaning of hieroglyphs. They're using a computer language developed in the mid-1990s to create a common Web publishing format.

Extensible markup language, or XML, is not a language in the proper sense, but rather a way of expressing information so that it looks the same no matter who's using it. Law enforcement agencies, for instance, might express a suspect's physical characteristics, such as height, weight and hair color, in XML. Today, many agencies express those characteristics using different abbreviations or they store the data in different computer languages. But with XML, each characteristic would receive a unique "markup tag," a kind of shorthand. The data then could be shared among many agencies because it would look the same in any computer format they use.

In September 2001, the General Services Administration led the formation of a federal working group to bring together technology project managers to identify ways XML could help agencies, says Marion Royal, the working group's co-chairman and the GSA's expert on XML.

The working group is building an XML registry. It's essentially a library for storing what Royal calls "fragments," common ways to express a name, address or phone number-basic pieces of data that many agencies would need and could share in the course of their daily work.

The working group's XML registry is still being tested. But the Defense Department already has built its own registry, and agencies such as the Environmental Protection Agency, NASA and the National Archives and Records Administration are participating in the working group's project, Royal says.

The group hosts monthly discussions, which attract 50 to 75 people, mostly project managers and contractors, Royal says. The group's Web site, XML.Gov, has become a key source for news and events throughout the government and in the private sector.

COMMON TERMS

XML is a perfect fit for federal agencies, says Tim Bray, the founder of software manufacturer Antarctica Systems and an XML co-inventor. Agencies hold onto data longer than most companies do, he says. For instance, the Social Security Administration maintains records on people for most of their lives. Often, the data doesn't change.

However, the same can't be said for the technology used to house or process that data. Some agencies store information in word processing documents, Bray says. The documents' contents remain the same, but new versions of word processing software are released every year.

Five years down the road, an agency that wants to share its data may be limited by how many other agencies have word processing programs that can read it. The same is true for information housed in databases and e-mail. But by expressing data in XML, any version of a software will work, so long as it accepts XML. Thus, the information always will be accessible.

Royal says most agencies recognize XML's potential, and now they're ready to test the waters. Many are intrigued by the prospect of using XML to deliver services through the World Wide Web, he says. The Education Department used XML to construct its e-Grants Web portal. As a result, the department can pass along educational grant information and applications to other agencies and to outside organizations and citizens.

The Census Bureau used XML to automate the creation of more than 650 types of forms in its most recent national economic survey. Each form pertains to a different industry, but many of the same questions are found on several forms. Census created a central repository for those questions, and tagged each one in XML. Thus, creating the forms was almost entirely automatic.

Law enforcement and intelligence agencies may have the most pressing need for XML, as new homeland security initiatives and the worldwide hunt for terrorists demand a seamless flow of information among many players.

To help law enforcement agencies develop an XML-based standard for sharing data about suspects, a nonprofit consortium of businesses and agencies is stepping in. The Organization for the Advancement of Structured Information Standards (OASIS) has formed a committee to examine potential XML standards for ways to express evidence of suspected terrorist activity.

The Justice Department had been working on a "set of vocabulary" to exchange information about suspects, Royal says. Through OASIS, the department would like to extend its work to include state, local and international governments. The group is recognized around the world, so if the law enforcement committee can find an agreeable XML standard, agencies will have gone a long way toward improving data-sharing.

SETTING STANDARDS

Groups like OASIS can help set standards for specific users, but ultimately the Office of Management and Budget may decide how XML will be used on a governmentwide basis.

OMB is creating a federal enterprise architecture, a governmentwide blue-print of how thousands of technology systems could work together. At the heart of the architecture is a series of reference models.

A business reference model, for instance, will detail technology investments that agencies have made based on the specific tasks they perform. A data reference model will identify common ways that agencies express information or common types of data in their systems.

That's where Royal hopes to find some new guidance. If the data model can pinpoint those common characteristics, they could be expressed as XML. Royal hopes OMB will release the data model in May.

XML advocates warn that the technology is no panacea for managing data. They agree that some better language may come along, but for now, XML suffices. "It's easy to understand, and it's easy to use, and there isn't anything better," says Bill Wright, president and chief executive officer of Computas, a technology services firm that's helping the Bureau of Printing and Engraving, the Census Bureau and the Environmental Protection Agency make use of XML.

Bray, XML's co-creator, says his language's strongest potential lies in the realm of knowledge management, the study of helping organizations categorize and access all kinds of data, from physical reports to e-mail messages to the expertise of their employees. By expressing data in XML, agencies can ensure it will be accessible decades from now.