Seeking Semantics in Government

Anyone who uses Twitter and has to cram thoughts in to 140 characters knows that technology doesn’t always mix well with “semantic meaning.” That reminds me of an old Hollywood story (here’s a version from Wikipedia):

Cary Grant is said to have been reluctant to reveal his age to the public, having played the youthful lover for more years than would have been appropriate. One day, while he was sorting out some business with his agent, a telegram arrived from a journalist who was desperate to learn how old the actor was. It read: HOW OLD CARY GRANT?

Grant, who happened to open it himself, immediately cabled back: OLD CARY GRANT FINE. HOW YOU?

WashTechWashington Technology magazine has a long (overly long) feature today about semantic computing, entitled “Open Government Looks for New Technologies.”  It has nothing to do with Cary Grant, but I have a few minor quibbles with the article (written by a freelancer from New York).

The premise is in the subhead: “Web 3.0 could help make Obama’s dream of government transparency a reality.”  The article goes on to give a basic – very basic – primer on semantic tagging and its potential application in government uses. Underline that word, “potential.”

Aside from the new Data.gov website’s use of minimal Dublin-Core metadata, there’s no actual government use cited. In fact, despite the premise, the article actually contains more evidence that government agencies are actively shying away from adopting semantic approaches. A spokesperson for GSA is typical, saying only that ““We are monitoring the situation as the technology matures; it is not factoring into our business requirements at this point.”  And a spokesperson for the site at www.Recovery.gov, now controversial for the manner in which it was contracted out, says they are “focusing on other priorities.” 

In fact the article entirely omits those few sections of government which to my knowledge are actually making good early-adopter use of semantic approaches: the intelligence community!  There are other efforts as well sprinkled across civilian agencies; the article’s author would have done well to quote one of the active semanticists in Washington itself (see here and here for some resources).

The real advances in semantic technologies are coming not in government, but in the commercial sector – of course.  (The article makes no nod to efforts like Twine or Bing’s use of semantic search from Powerset and other research.)  We can all agree that it is early days yet for semantic technologies in government use. Articles like this one pop up periodically promising great things; here’s one from back in 2006 with the same basic message: “Semantic Web Ready for Prime Time.”

Perhaps it’s too much to hope that tech-publication editors maintain a bit of historical context the next time someone pitches a breathless article hyping “new technologies.”

In fact, the editor might have noticed that, directly contradicting the entire premise of the headline and intro paragraphs, the money quote comes right at the tail end of the article, with one of Washington’s most prominent open-government advocates poohpoohing any near-term semantic focus entirely:

The time and effort required to tag and describe the government’s vast data holdings represent another adoption challenge. Clay Johnson, director of Sunlight Labs, expressed concern that the government might become preoccupied with formatting data rather than releasing it. Sunlight Labs is an open-source development team that launched as a project of the Sunlight Foundation, an open-government advocacy group in Washington. “I would hate to see them get bogged down in trying to make their data Semantic Web compatible before it even sees the light of day,” he said.

Oh, and one more minor quibble: the full subhead of the article reads: “Web 3.0 could help make Obama’s dream of government transparency a reality, but he’ll need a second term to see it happen.”  As of today, President Obama is beginning only the second six-months of his first year of his first term. An increasing proportion of the country appears already to believe he’s not doing too well

Let’s not get ahead of ourselves with planning for the second term’s agenda, before making progress on the first.

Email this post to a friend

AddThis Social Bookmark Button

2 Responses

  1. Yep, not a good article.

    Innovation often takes place in unexpected ways. Even in government.

    Motivated by Kundra and Chopra, the sentiment in civilian agencies is more amenable to new technologies in this administration than the last. The good news is the folks at RPI have already translated data.gov datasets into RDF.

    http://data-gov.tw.rpi.edu/wiki/Main_Page

    Stay tuned and let’s hope the meetup community can take a lesson from the good work at Sunlight and make use of the datasets Hendler & Co have provided.

    Rick

    Like

  2. Lewis, thanks for the tweet and the links to your other online musings. Reading some of your older posts now.

    Have you come across these folks? http://www.insilicodiscovery.com Interviewed them today. Good traction already in pharma and banking–extending the spreadsheet paradigm to include a semantic base and expand the integration capability. They design an OWL model for each customer, and tune it as customers feed spreadsheets into the integration mix.

    The result accelerates and simplifies the ability to update reports, and increases the visibility execs have with the higher level of aggregation each report is endowed with.

    Philosophy is that the exec suite is familiar with the spreadsheet interface and it’s more sophisticated now, so why not build under it? Very smart way to introduce the power of ontologies under the covers directly to the execs. All the customers see is the updated reports in the spreadsheet form they’re familiar with. The SaaS updates the OWL model as necessary. Just thought you’d be interested.

    I followed @amcafee’s pointer to your most recent. Great conversation starter.

    Like

Leave a comment