Federal IT projects always seem to be ineffective, expensive, and slowly done. On the other hand, projects using new software development tools and approaches seem to get the job done fast and cheap.
For example, the Coburn-Obama bill mandated that the OMB build a database of government grants and contracts. Sunlight Foundation funded a nonprofit, OMB Watch, to build a similiar database, took 'em eight months and $320,000.
The Coburn-Obama bill allocated $12 million for it.
Turns out OMB purchased the OMB Watch software, which is the basis for USASpending.gov.
What do we learn from this?