It's one thing to build a data standard. Getting it implemented is a whole new ballgame.
The Government Accountability Office (GAO) has taken a long look at the 57-element data standard prepared by OMB and Treasury to enable the DATA Act. GAO identified 13 best practices for data standards and found all the data elements in the set met at least 9 of the criteria, with 12 meeting all the criteria. But they found some real problems in there.
For example, there's an entity called Primary Place of Performance. You need to be able to say where money is being spent. But the definition includes the phrase “where the predominant performance of the award will be accomplished”. What if the project is a road? You could interpret the Primary Place of Performance as the town hall or state capitol, depending on whether the road is considered local or statewide. Or you might think it's the contractor's worksite, or the contractor's business address. Or you might have a geolocated mileage marker close to the project and use that.
Developers can interpret ambiguous definitions like this in ways that make perfect sense for their systems and their stakeholders' needs – but which undermine the ability to share data across agencies. So data standards bodies need to issue guidance about how the standards are to be interpreted. They also need to invite feedback and experience reports so new use cases can be tested against the standard.
It's hard enough working against organizational inertia, stressed budgets, and not-invented-here syndrome. Get the guidance out there if you want to get implementation. And accept that every standard will need to adapt to changing needs and growing awareness. GAO
Comments