In an effort to further develop the use of XML in the public sector, the Center for Technology in Government, a research organization at the State University of New York at Albany, walked the walk: It converted its Web maintenance process to XML.
The center last year shifted from static documents, which needed to be reformatted each time they were updated by writers, layout designers, or programmers, to open standards that let one version of each Web page be automatically repurposed to any format. "Before we converted to this technology, we had one full-time person managing a site that was 1,500 pages, and that was all this person was able to do," says Derek Werthmuller, the center's director of technology services. With XML, the site is 10 times larger, and the center only spends one day a week managing it. "That means content is coming into the site more quickly, we can write more reports, and we can run more projects at the same time."
The center's extensive return-on-investment analysis of the project revealed that using XML cut in half the personnel costs of running its Web site. "Instead of struggling under the weight of our Web site, we have this opportunity to branch out and do projects we otherwise wouldn't have time to do," program manager Donna Canestraro says. "We're using our experience to help agencies look at the benefits of using XML in government."
No doubt, XML is an appealing option for cash-strapped states and federal agencies because it's cheap, easy to manage, and provides a bridge between incongruous data sources to foster open communication both within and among agencies. And it's taking off on many levels. Most government bodies have interspersed XML into their overall technology enterprise to some degree. At the Library of Congress and National Archives, for example, XML is the standard for digitally storing records. What's more, it offers freedom. "XML gives us the best chance to own our own data, because it's not in a proprietary format so we're not beholden to a vendor," says Mike Short, chief of enterprise development at the New York State Department of Civil Service.
But such liberty doesn't always come easy. Just ask Peter Quinn, the former CIO of Massachusetts. In 2005, Quinn and his team were tasked with solving Massachusetts' data incompatibility problem and its inability to share data stemming from years of technology being deployed in silos across the state's executive branch. Quinn also wanted to ensure that technology never separated public from government information. Quinn's solution: OpenDocument format, an XML-based file format for office applications sponsored by vendors such as IBM and Sun. The format was standardized last year by the Organization for the Advancement of Structured Information Standards.
"OpenDocument format, and applications that read it, are far more likely to be available in 300 years when our great-grandchildren want to read the electronic records we create today," Quinn said in testimony to the Massachusetts Senate Post Audit and Oversight Committee in October. "Furthermore, multiple office applications support ODF even today, so citizens making public records requests will have a choice of office applications when they read electronic public records created in that format."
Quinn's team mandated that all executive branch documents be created and stored in OpenDocument format by 2007. That decision wasn't well received by Microsoft, which stands to lose millions of dollars in sales of its proprietary Office licenses. Though Microsoft is working on its own open-file format called Open XML for the 2007 version of Office, it won't be ready in time for Massachusetts' adoption plan. What's more, Microsoft will be the sole proprietor and developer of this format, whereas the OpenDocument format is a collaborative work by vendors.
The firestorm flared in the political arena as politicians aligned with either Quinn or Microsoft. While the OpenDocument format mandate is still in force today, the controversy led to Quinn's resignation at the end of last year (see story, Mass. IT Chief, Supporter Of Controversial Open Software Format, Quits). Speculation remains that Louis Gutierrez, Quinn's replacement appointed last month, will continue the initiative.
Politics and IT just don't mix well for California CIO Clark Kelso.
Photo by Jeffery Newbury
Kelso admits the political saga behind the Massachusetts initiative has entered his mind. "It makes all of us pause for a minute and pull back a little bit, and that's not a good thing," he says. But he refuses to allow politics to outweigh his IT decisions that are based on comprehensive analysis of what technology best meets a business need. "We're in the public sector and we live in both choppy and, at times, shark-infested waters. I'm not going to let IT issues be framed in a political context," Kelso says.
Kelso uses open standards where they most make sense, such as adopting a service-oriented architecture that embraces XML. But for the most part, adopting one standard doesn't make sense for California, where state agencies are large and tend to make their own IT decisions.
Too Powerful To Ignore
XML tends to move into government most successfully at the agency level. It's far too powerful a tool to ignore, says New York's Short. Last year, he spent six months participating in an XML test bed project that the Center for Technology in Government sponsored aimed at duplicating the center's success in using open standards for Web development and content management. Based on his experience in the project, Short is using XML to transform both the process for developing and administering examinations for state and municipal jobs and to post qualified candidates for positions in a way that state agencies can easily access.
Today, those exams and postings are created in Word documents by one person, then shipped to a Web publisher, who translates them to HTML so they can be posted on the Web. That person transfers the documents into Corel Ventura conversion software to be converted by someone else into a PDF format. There's room for error at each touch point, Short says. Plus, if any person along that way finds an error, there's a chance that person won't be able to find the other people in time to make the changes in other formats.
With XML, documents are only published once and can be viewed by any application. What's more, an error only needs to be corrected once and the changes are automatically made throughout any other version of the document. In addition to risk prevention, XML offers efficiency payoffs such as eliminating the need for a Web publisher.
Short's long-term dream is to use a standard language for all of the Civil Service Department's exams to simplify the examination-building process. Today, exam builders have more freedom of language in how they ask questions. Short's plan is to use a standard language and XML so that test creators simply check the boxes for questions to be added. "We can have radically different exams run by the same program with the XML data driving the examination," Short says.
For some agencies, XML is a long-standing and pervasive technology that has proved its worth in cost and time savings. The federal Bureau of Justice Assistance, a division of the Justice Department, has been working on an XML-based standard, the Global Justice XML Data Model, to promote information sharing across all local, state, and federal justice agencies for the past six years. The bureau has collaborated with federal, state, and local agencies to build on the existing XML development standards, and then keep those results transparent so that other states or local agencies can use them.
The bureau's XML schema is so well developed that it can implement large national information sharing projects, such as linking states' sex-offender registries and national law enforcement databases to be published for the public, in a matter of months at a low cost, says Patrick McCreary, associate deputy director of the Bureau of Justice Assistance.
Similarly, state law enforcement agencies are able to use the schema to quickly develop their own programs. Missouri used the code to complete a project unifying court data from 49 unique court-case management systems into a single database, the largest statewide implementation based on the Global Justice XML Data Model. The project will be used as a model for other states to follow. "Keeping the data model open promotes reuse whenever possible so one jurisdiction can leverage the work another has done and adapt it to whatever their needs are," McCreary says.
But the only way that strategy will work is if the people involved in developing them know what they're doing, McCreary says. That's why his department has trained hundreds of developers in both the public and private sectors on how to use the standard correctly. "With a consensus standard, unless you have a training practice so people are implementing it in a similar manner, they won't be able to share information effectively," he says.
The CTG understands that. To help ensure that XML is being used consistently as a Web publishing tool, the center has collected and analyzed all the data from its test bed project and by the end of the month will post a free resource library on the Web. "We've come up with recommended practices, lessons-learned documents, an XML library of tools and stylesheets, and an XML starter kit," Canestraro says. The center's extensive ROI analysis from its own XML project is also available for public viewing.
Whether used as a Web publishing tool or a document format to save money or improve collaboration, open standards are rapidly weaving their way across every level of the government technology enterprise. And the open-standard controversy is simply a page in a long and continuing history of IT and government, California's Kelso says. "Next year, something better than XML may come along, but my business focus will be the same," he says. "I expect our technologists to take the next greatest thing and apply it to our business needs."