After Years of Pushing, Senate Finally Releases a Custom XML Blend of Roll Call Votes
BY Nancy Scola | Wednesday, May 6 2009
It's ask and ye shall receive XML, it seems. After years of pushing by open government advocates, it took just a "Dear Colleague" from South Carolina Republican Jim DeMint and a gang of bipartisan senators to finally spur the Senate Rules Committee to, this week, change its longstanding practice and begin publishing votes in a structured data format that makes for easy analysis, visualization, and mashing up by the public. Rules Committee chair Chuck Schumer (D-NY) perhaps saw that, with transparency on the march in Washington, this was an inexpensive, easy win. Almost immediately, the Senate began publishing a comprehensive XML list of roll call votes and details on each bill's vote, also in XML.
A recent Politico article raised the specter that senators were standing in the way of easily accessible roll call votes because of a desire to serve as an interpreter (or, to be less gentle about it, spinner) of those yeas and nays for the public. DeMint and his allies weren't having any of that, writing "The public is capable of interpreting our votes on its own." There also may have been a bit of friendly bicameral rivalry at play. In their Dear Colleague, the transparency-seeking members bemoaned the fact that "the Senate has fallen behind the House of Representatives in terms of vote transparency."
The Senate was inspired by the House's XML example, perhaps -- but not enough to let the lower body dictate how the Senate is going to structure its data. Despite the fact that the House has been in the business of publishing roll call votes in XML -- by default -- for years now, the Senate whipped up its own format without consultation with the other body. As a result, Senate and House votes are now made public in significantly different formats. The House identifies each voting member using a <legislator name-id> tag that connects up with Congress' own Biographical Directory, the standard resource used to ID members of Congress. The Senate, on the other hand, uses an <lis_member_id> field, used, it seems, only for internal Senate business.
Those data format idiosyncrasies, though, will only be a minor speed bump for open government advocates who now have their hands on a new cache of data.