Bridging the gap between federal, state and local agencies
Barriers to web service implementation
Three big challenges must be overcome before a vibrant web services network can be created. First, security and confidentiality must be managed; second, data definitions and terminology must be agreed upon and standardized; and third, a general reluctance to share data must be overcome.
Managing security is always a concern, Sweden says. “That is a requirement when we talk about information sharing—maintaining security, privacy, authentication…. Along with that information going from agency to agency, we’ve got to be sending information about authentication and authorization: the metadata regarding security and privacy.”
Still, there is always the risk that data will fall into the wrong hands whenever it leaves the home office. This is not unique to web services, but is inherent in all data sharing. “Opening up information carries risk,” says NASCIO’s Eric Sweden. “You will never get to zero risk.” But, Sweden adds, “Information that is not shared cannot be used. Think of information as an asset. Are we getting full value out of the asset?”
Agreeing on data definitions can be as daunting as wrestling with confidentiality concerns. If you think concepts like “address” are simple and straightforward, read the Federal Geographic Data Committee’s “Street Address Data Standard” (http://urisa.org/node/227). Be sure to read and memorize all four parts.
Finally, many public agencies are unsure of their obligations to provide or withhold access to data. The Freedom of Information Act (the federal statute and its state-level counterparts) is generally vague about the obligation to provide web access to digital records. Some agencies struggle with HIPAA (Health Insurance Portability and Accountability Act of 1996) and counter-terrorism guidelines that restrict the release of some data.
A note on GML and PMML
XML’s great strength—it’s almost infinite extensibility—also dictates that to be useful, XML schemas must be agreed upon by all parties in an exchange network. That’s why major IT industry players have banded together to hammer out specific XML standards for specific purposes. Two of these are Geospatial Markup Language (GML) and Predictive Model Markup Language (PMML). GML allows geospatial data to be transported as XML, while PMML does the same thing for statistical and data mining models. Both these technologies are maturing to the point where we are starting to see applications in production. Either could be the subject of dedicated articles, but I at least wanted to mention them here.
The time is now for real-time data linkages among public entities, and real-time data feeds to the public. As long as sufficient attention is paid to issues of standards and security, data sharing becomes relatively simple and sustainable. Government efficiency, accountability and performance can be enhanced, while the public’s role in deriving meaning from the available data can be encouraged.