Is there a low cost shortcut to make government data more connectable without having to tackle the barriers of a full linked data or rdf approach?

In order to make government transparent, we must make government data sets able to connect to each other.  One example of data which needs to be connected is  connecting policies to procedures to efforts in government. This is the DNA of government and this information is needed to truly innovate government.  But there are many other examples of government datasets both kept internally and released publically which would benefit from connectivity standards.  For instance just being able to find data which applies to a specific office in an agency or a specific program.

Problem: How do you increase the connectivity of government data so it can evolve into government information?

And is there a low barrier and low cost path to get there?

Data becomes useful and can start to spur innovation when it is connectable to other data.  One way to make information more connective is to have field definitions called XML tags which indicate the type of information in that field.  So if we have a tag called <agency> and that field contains names of government agencies, then you should be able to connect that information with other data sets about government agencies.

But there’s a problem with having common field names (data standards)!

Unfortunately even though everyone may be agreed on the name of the field or have a data standard, they may not agree on how to refer to specific agencies.  In short it doesn’t have a connectivity standard. So data in these fields might different between different data sets.  It may read “EPA” or “Environmental Protection Agency” or the “US Environmental Protection Agency.”   So even if 2 data sets shares the same XML language, it may not be connectable easily as a computer doesnt necessary know that EPA=Environmental Protection Agency= US Environmental Protection Agency.

A solution borrowed from Linked Data or RDF?

So what we need is a consistent way to name things in the world.  Linked or Semantic Data has a concept called a Universal Resource Identifier or URI.  A URI is simple a URL or web page address which is permanent and serves as the definition for an object or concept.  This is similar to the concept of having a primary key in a database, but in this case the database is the internet.

How do we get started?

I know what you are thinking.  This is another major IT project for which we don’t have funding.  Or this means that web managers will have to rename pages and then keep then permanently which would be a huge problem.

It’s easy. Make a it a redirect!

To avoid the problem of adding to the complexities of maintaining government websites, we would use a redirect service to serve as the repository of the URLs.  In RDF parlance, these are called  PURLs (Persistent Uniform Resource Locators).   So that there would only be one database of the URIs but that would not have to contain the definition pages, it would simply redirect to a web address.  So if a new content management system is installed, then we simply update the central URI registry.

Gee, if only we had a low cost redirecting service which could handle high volume!

Well we do. It’s called Go.USA.Gov.  The purpose of this redirecting service is to offer a URL shortening service for federal agencies. This is needed and a great idea. I propose we give this service a dual use. We add a subdirectory “/uri” and allow people to choose a URL name instead of it always being shortened.  This is similar to how the commercial service tinyurl.com works.

Or for slightly for effort…

We could copy the existing mode, make the custom name modification and change the name to URI.gov. This won’t require any additional license purchase since the code is open source and most likely could still be placed on the same server and use the same resources as Go.USA.gov.

But who would make up all of the URIs?

This is where the crowdsourcing aspect comes in. You simple setup the service and standards for how to do this and you evangelize the potential for greater transparency to web managers and the CIOs.

But it won’t be perfect!

Yes it won’t. As is the case with other crowdsourcing efforts. But it is way to solve 90% of the problem for .1% of the effort. Once it gets going and the value of the URIs becomes clear, then it can be officially be incorporated into federal procedures by the CIO council or web managers council.  We can start to use service then to connect Policy to Procedure to Effort as well as make connections between other useful datasets about government. This not have the barriers of taxonomy and technology which a true linked data strategy has, but provides some of the benefits of the linked data model.

So cheap, quick and potentially very useful.   Why not?

Note to Linked Data guys: Guys this is not to say we should not move toward rdf or rdfa but instead is a way to show the power of one of the benefits of it.