Big News!

To kick-start 2010, we have updated virtually every tool we have in the OpenCalais arsenal. Check them out and let us know your thoughts.

• Version 4.3 of the OpenCalais service features enhanced Social Tags for categorization; improved disambiguation; new entities, facts and events in English; and an entirely new feature called News Names.
• By popular demand, the improved Tagaroo plugin for WordPress gives users more control over the tagging process.
• Our URL submission tool, SemanticProxy, has been normalized to ensure that its results are identical with those you would get by programmatically submitting content to OpenCalais via the API. Additionally, we are now supporting JSONP. You can now specify a callback function when requesting results in JSON.
• Our Gnosis plugin for Firefox now includes categorization of processed Web pages in the sidebar and can sort entities by alphabetical order or according to frequency or relevance.
• And finally, we have published the OpenCalais Schema in OWL, enabling OWL enthusiasts to apply ontology based tools on OpenCalais metadata.

Below please find further details on OpenCalais 4.3, Tagaroo and SemanticProxy.

New in OpenCalais 4.3

Improved ‘Social Tags’: We are expanding on our popular Social Tags categorization technique by adding more generalized, aggregate tags.

For example, if a blogger is comparing the racing performance of sports cars like the Ferrari 308 GTB and Porsche 959, OpenCalais 4.3 will suggest auto racing and motorsport as Social Tags, in addition to the more obvious sports cars.

NEW! ‘News Names’: We are instituting a process of name normalization that represents a first step toward our more robust vision for person disambiguation. Whenever a partial or extended name appears in content, OpenCalais 4.3 will return the names it finds as usual, but will now also suggest the most commonly used form of that same name.

For example, for articles containing Barack Obama, Obama or Barack Hussein Obama, OpenCalais will suggest not only the partial or extended name it found, but also the more frequently used Barack Obama.

New Entities, Facts and Events in English, including:

• New Natural and Manmade Disaster attributes that reveal these disasters’ effects
• Supporting data for upcoming events that will enable OpenCalais to recognize new Movies, Music Albums, etc., as well as anticipated Medical Treatments
• More Political Events and new items such as Diplomatic Relations, Political Endorsements, Poll Results and Voting Results
• Enhanced Person Career extraction that includes political party affiliations where those are included in the text.

The 4.3 release also features improved Simple Format and Microformat outputs, as well as several extraction bug fixes. For technical details, please see the full release notes here.

Tip Top Tagaroo!

By popular demand, we have also made two key improvements to our Tagaroo plugin for WordPress.*

Tagaroo no longer suggests tags while you type (which many users had found disruptive). Instead, you simply click a button when you are ready to see tag suggestions, and then select the tags you want.

We also added the ability for you to select / highlight a portion of text in your post and get tag suggestions for that text alone, in the event that you don’t want to tag the whole post.

*Note: You must have a hosted site – or your own server – where you have access to install WordPress plugins. Blogs hosted on WordPress.com won’t work with Tagaroo.

SemanticProxy Support for JSONP

Finally, we have also made two important changes to our URL submission tool, SemanticProxy.com.

First, we normalized the service to ensure that the results it returns are identical to those you would get by programmatically submitting content to OpenCalais via the API. This entailed discontinuing the HMTL cleaning that SemanticProxy did on its own, which had occasionally caused the service to return false content-type errors.

Second, we added the ability for you to specify a callback function when requesting results in JSON format. Callbacks are useful for web service requests in client-side JavaScript and provide a relatively simple way to invoke web service requests across domains.

To use it, specify the callback function name followed by the JSONP output format. For example: http://service.semanticproxy.com/processurl/licenseid/jsonp:processresul...

The JSONP output format wraps resulting JSON output text in parentheses and the provided callback function name. For the request above, wrapped results would be: processresults({"doc":{"info":...json output...”} );

As JSON is native to JavaScript, you can access the elements inside the returned output, similar to passing a JSON object reference to the processresults function. In other words, if your JavaScript has a processresults function which takes a JSON object as input and manipulates it, this function will be invoked automatically.

Please see the attached demo html to see this in action.* 

*Note that the demo will not work in IE. Also, we have restricted the volume for this demo key.  If it does not resolve, please plugin your own OpenCalais API key.

Jump in!

Let us know what you think about these updates. We appreciate your feedback and look forward to any suggestions you may have.