Owen Ambur | The data landscape

'I like to call people's bluffs,' Owen Ambur once told us. Anyone who has attended meetings of the CIO Council's Extensible Markup Language Community of Practice can attest to his insistent nature. As co-director of the group, Ambur tirelessly encouraged XML-based interoperability. In January, Ambur retired as a senior architect at the Interior Department, where, 20 years ago, he was one of the first managers to recognize the value of electronic recordkeeping. We caught up with him in his Silver Spring, Md., home to discuss the importance of recordkeeping and StratML, an XML-based schema he is developing, with Adam Schwartz of the Government Printing Office, for agency strategic plans. I came to Washington in 1973 from South Dakota with a newly elected congressman, Jim Abdnor. The bane of the existence of a congressional staffer at that point was correspondence. That was before the days of computers and the Internet, so it was paper correspondence. We used manual typewriters without correcting tape, and I was a lousy typist. So I set up a form [response] letter that one of our clerks could run out on the automatic typewriter. So we got an immediate response back to the constituent, and then as soon as the agency responded, we'd transmit that response back.After eight years in the House of Representatives, Abdnor ran for the Senate and won. We moved into the Senate. The Senate administrative office previously did the requirements analysts for office automation system for the entire Senate. So we could select one of three minicomputer systems. We got a turnkey system with word processing, spreadsheets, file-management capability. Minicomputers were workgroup machines ' that's what they were designed for.And this was in the early 1980s. And so one of my frustrations over the last 20 years is that the bureaucracy that I had been in is not managing information as effectively today as we were in Senator Abdnor's office in the mid-1980s. The real benefits of records management is the day-to-day management of activities of the agencies themselves. We all spend countless hours trying to find information that should be right at our fingertips.If we had good records management, fraud would be nonexistent. Fraud relies upon poor recordskeeping systems. To the degree we agree to use electronic information systems to conduct business, we would have nearly perfect records of those transactions.To me, by far the most important reference models are the data reference model and the technical reference model. The technical standards should be developed by voluntary consensus standards organizations, but the records required to conduct we-the-people's business are inherently governmental. The data reference model was the last of the models to be produced because it was, from the government's perspective, the only one that matters. If agencies don't understand their data architecture, they don't understand their own business. Over the years, I began to think that maybe I'm the one who doesn't get it. These are not bad people. They're not evil. They're not dumb. Maybe they just don't want to do this. Maybe there are deep underlying human motivations that mitigate against good records management. And, if you look, you find an incredible number of motivations [for not setting up a recordkeeping system].The first is the failure of human memory. Daniel Schacter, [a Harvard professor who] wrote a book called the 'Seven Sins of Memory,' said what it comes down to is that we don't like to be confronted with evidence that conflicts with what we think we know. But what we think we may know may or may not have much to do with reality. As Schacter points out, our memories are not photographic. They're reconstructive. We put things back together based on not only what actually occurred, but everything since then and our current emotional state. It's not good or bad. It's just human nature.[Another reason] why people don't want to have good recordkeeping systems is that to the degree that the record is unclear, the power of the [individual] is increased, because it gives them greater latitude to do whatever they think is right. I don't think the vision is clear enough. I'm not attributing it to any ill will. I think the major factor, aside from these implicit motivations people have but have not yet acknowledged, is the complexity of it.But it is not so complex that it cannot be comprehended and addressed in small, manageable chunks. Frank Raines told us this 11 years ago, when he was director of the Office of Management and Budget. One of 'Raines' Rules for Federal Information Systems Investments' [An OMB memo about system procurement. See GCN.com/821] was to address IT in small, manageable chunks. It's an idea that agencies routinely ignore. StratML eventually will enable literal linkages between strategic goals, objective statements and any other record the organization creates. If you find a strategic objective that is important to you, you should be able to follow that strategic objective to links to any other records that are supportive of that objective.If agencies were truly serious about strategic alignment, why would their Web sites be anything other than their strategic plans? By definition, it means they are doing something other than aligning with their strategic plans. That is not to say strategic plans are the be-all and end-all, but if there is a gap between the strategic plan and the Web site, maybe the strategic plan should be altered, or maybe what [the agency is] doing should be changed.