Government seeks accord on XML

Growing support among federal agencies for the increasingly popular data exchange format raises concerns that things might be moving too fast.

Margaret Kane Former Staff writer, CNET News
Margaret is a former news editor for CNET News, based in the Boston bureau.
Margaret Kane
5 min read
The federal government isn't known as a pioneering early adopter. But growing support within U.S. agencies for the popular XML data exchange format has raised concerns that, for once, things might be moving too fast.

And the government's purchasing arm is now studying whether a central directory should be built to prevent Extensible Markup Language (XML) from creating the sort of integration headache it was intended to solve.

The General Services Administration has commissioned Booz-Allen & Hamilton to study the creation of a registry and repository for Extensible Markup Language, which has become widely popular in the business world, where it is seen as a way to simplify data exchanges between disparate businesses and software programs.

XML lets programmers create definitions for any type of data that will be handled, such as names, product ID numbers and lot sizes, so that computer programs can instantly recognize the information being transferred and handle it properly--as long as both sides agree on definitions.

The language has also been a hit with the government. With a budget of more than $48 billion--that's 16 times larger than the annual IT expenditure at General Motors, for comparison's sake--the federal government is easily the largest consumer of computer hardware, software and services on the planet.

But the fear is that, without a formal policy, multiple agencies will come up with conflicting definitions and that a technology aimed at making networks and computers communicate more efficiently will end up making them incompatible. The same XML incompatibility fears afflict private sector companies, on a lesser scale. It's the government's vast scope that makes it especially ripe for software communications snafus.

The GSA study comes after a recent General Accounting Office report questioned whether, without a firm government policy, the use of XML could build out of control.

"Although agencies need the flexibility to tailor XML-based systems to meet their unique needs, they risk building and buying systems that will not work with each other in the future if their efforts do not take place within the context of a well-defined strategy," the GAO report stated.

As one of the authors of the report put it, "If everyone develops their own data, schemas and definitions, we have a Tower of Babel and we haven't gotten what we wanted."

The registry and repository will help by providing a central location for agencies working on XML to list their projects, and to store data definitions used frequently by governmental agencies. Other agencies would then be able to find the data and compare that to the work they are doing.

Not limited to government
Of course, concerns about incompatible XML definitions exist in the private sector as well. To head off compatibility problems, companies in a given industry, such as insurance or financial services, get together to agree on a standard for that industry. But the concerns are somewhat heightened for government because of its sheer size and basic bureaucratic nature.

The federal government spans multiple industries, and must work with outside businesses and contractors, as well as communicate with other agencies with the federal system. So the problem is not just making sure that, say, engineers for one federal agency can exchange documents with private sector engineers, but making sure that they'll be able to communicate with engineers in other branches of government as well.

Gartner analyst Rita Knox says XML standards efforts should minimize the amount of "hard coding" they incorporate and should maximize flexibility.

see commentary

Those concerns led to both the new GSA study and the GAO report, said Marion Royal, agency expert at the GSA and co-chair of the XML Working Group of the federal CIO Council, which had established its own registry as a "proof of concept."

"We decided to conduct a business analysis to see if this made sense on a governmentwide scale," he said.

Starting from the results of the GSA study, the council will work with the Office of Management and Budget to build plans for an official registry and repository into the fiscal 2004 budget. An interim report is due at the end of July, with a final recommendation coming in October, Royal said.

Meanwhile, the council may try to set up a larger project using off-the-shelf software or by working with another agency.

What needs to be done, some argue, is to work on the problem from both ends: develop a set of best practices and standards that will apply across agencies, and encourage individual developers to start working on projects and setting up registries where their work can be shared with others.

"Our concern is at the macro level," said John de Ferrari, assistant director of the IT team at the GAO. His team produced the XML report earlier this year.

"The technology path is there, or it's on its way, and if you want this to happen you'll have to set standards and (conduct) the kind of cultural change needed to make it happen," de Ferrari said.

That notion is backed by those that work with the federal government on XML-related technologies.

"The highest value for all of us is that all the agencies work together so they have common DTDs (document type definitions, or sets of rules for XML data tags) for common things," said James Shay, CEO of First2File, a company that makes patent filing software. "Certainly there's some things that are specific to patents, but there are other things that are not unique. As software providers...we'd like our software to be compatible with every agency we work with."

Working from the ground up
Some of the bottom-up work has already begun. The CIO Council, an interagency forum for government technology chiefs, chartered the XML Working Group, which has been studying best practices and standards. The group distributed a draft list of federal government-specific best practices in January, and has been accelerating the creation of a registry for "inherently governmental" data definitions and schemas, said Owen Ambur, a systems analyst at the U.S. Fish and Wildlife Service and co-chair of the group.

The group has also composed a list of some of the XML efforts already being undertaken by different federal agencies. More than 30 are listed at the XML site, covering topics ranging from child support enforcement to federal grant processing.

"There's way too much policy that has been written that can't be enforced. It's more important that we provide the tools that will support the policy, and enable agencies to do the right thing, rather then just tell them to do the right thing," Ambur said.

And the government has already begun working with standards bodies to make their needs known. For instance, agencies including the departments of Justice, State and the Navy are working with the Organization for the Advancement of Structured Information Standards, or OASIS, a consortium developing electronic business standards including XML.

But there is strong support for developing at least some sort of over-arching government policy. The GAO report was written at the request of Sen. Joe Lieberman, D-Conn., who has also sponsored a bill that would establish an Office of Electronic Government under the OMB to promote interagency collaboration on systems and standards.

"I view this as a two-step process," OASIS Chief Executive Patrick Gannon said. "The government needs to actively participate with industry, preferably with industry associations, to help put their requirements on the table. And then they need to participate in doing pilot implementations so they gain experience."