As part of a recent leadership program here at education.au Simon (no blog), Will (no blog) and myself took the opportunity to develop a concept for a new style of search engine, designed to supplement the edna distributed search engine with better support for querying structured metadata.
We designed a new type of search interface, using query refinement to guide users to their goal, rather than the traditional advanced query form. This was accomplished using the open source Solr search engine (hence the name - Johannes Kepler discovered the laws which govern the movement of the planets in the solar system. Yes… we know it’s kind of a geeky joke.)
Using the OAI-PMH protocol we were able to harvest metadata from a number of open repositories around the world.
In our test system we ended up with 2.2 million records from around 40 repositories. Thanks to Solr we’re still able to do sub-one second searches on most queries.
(please visit website if this movie doesn’t work in your blog reader - or if you don’t see the movie at all)
We have implemented KeplR as a Javascript application, which means it can be places on any third-party site with a single line of Javascript. It is also possible to restrict the search to a single repository - or combinations of repositories - which means that site specific search engines can be implemented with no coding. I’ll save a demo of that for another post.
Meanwhile, you can play with our prototype system at keplr.uat.educationau.edu.au
Post a Comment