Census: Going Back to Paper Due to "Lack of Communication"


The U.S. Census Bureau announced yesterday that it was reverting back to paper from its plan of using handheld computers for the 2010 Decennial Census. The reason?

According to Director of the Census Steve H. Murdock's testimony before the United States House Appropriations Subcommittee on Commerce, Justice and Science, "the problem with the FDCA (Field Data Collection Automation) program was due to a lack of communication between the Census Bureau and the prime contractor for FDCA, and to difficulties the contractor had in developing the full scope of the project within our deadlines. From the beginning, we did not effectively convey to the contractor the complexity of census operations, and the detailed requirements that needed to be fulfilled in order to complete the operations that FDCA covers."

In U.S. Secretary of Commerce Carlos M. Gutierrez testimony, he said that, "In 2007, the Address Canvassing dress rehearsal was conducted, at which time development and scoping problems emerged. Reports from the Census Bureauâ''s field staff, consultants from the non-profit MITRE Corporation working for the Bureau, and the Government Accountability Office confirmed these problems. The departmentâ''s Inspector General also raised concerns."

"In late 2007 and early 2008, more than 400 new or clarified technical requirements were identified by the Census Bureau. Upon the realization of the large scope of requirement changes, Census Director Murdock established the 2010 Census FDCA Risk Reduction Task Force, to begin to propose and evaluate options to keep the FDCA program on track. These efforts served to clarify the issues and confirm the urgent need for action."

The action was to punt (and bad mouth the contractor as much as possible even as the Bureau "accepted responsibility").

Gutierrez's testimony, as damning as it is, fails to mention that both technical and management issues with the handhelds were raised well before May of 2007 - all the May 2007 dress rehearsal did is to confirm them. Even as the sirens were going off that major trouble was brewing and that urgent action was required to be taken by July 2007 at the very latest, the Census Bureau and especially Gutierrez himself kept their collective heads in the sand, all the while claiming the project was moving along smartly, and that the critics (like me) were unjustifiably bashing the program.

Guess I wasn't, after all.

So, another $2.2 billion to $3 billion will be spent on top of the $11 billion already allocated to complete the census. What the hell, it's only taxpayer money.

Kudos to the Census Bureau for creating yet another case study on how not to manage a large scale software project in government. A classic IT blunder and debacle all rolled into one.

That said, let's hope that Congress demands a thorough, open and detailed analysis of this project be under taken now - before the files are "lost" - and a plan developed outlining how the Census plans to automate the 2020 census actually using for once the lessons learned.

Want to bet that this won't happen?


Risk Factor

IEEE Spectrum's risk analysis blog, featuring daily news, updates and analysis on computing and IT projects, software and systems failures, successes and innovations, security threats, and more.

Robert Charette
Spotsylvania, Va.
Willie D. Jones
New York City