On July 1, the Ohio Public Library Information Network launched a new interface for searching the Ohio Web Library database collection at www.ohioweblibrary.org.
When OPLIN introduced the WebFeat OneSearch feature several years ago, librarians were happy. Finally, there was a nifty new way to search all of the statewide reference databases at the same time, instead of searching them one by one! It was indeed a great improvement.
By last fall, however, when OLC conducted librarian focus groups around the state on behalf of OPLIN, it was clear that OneSearch was not so nifty any more. Librarians had legitimate complaints about the speed of the search and the way search results were grouped. We repeatedly heard the word "Google"; librarians needed a search as fast as Google, with good relevancy ranking of the search results.
WebFeat offered some new products that would have been an improvement, but OPLIN decided to pursue an open source replacement for OneSearch, to give us the ability to continuously change and improve the search tool.
We elected to build a search page at www.ohioweblibrary.org that handles public access to the information databases provided by the Libraries Connect Ohio partners (the "Ohio Web Library" collection) and the OPLIN-hosted information products such as "What Tree Is It." The information intended for libraries remains on the oplin.org page, which is much cleaner as a result of the removal of the Ohio Web Library and other "public" information.
The underlying engine that drives the new Ohio Web Library search is open source metasearching middleware called "pazpar2" from Index Data. For those of you who are technically inclined, there is much more information about pazpar2 at indexdata.com. All you really need to know, however, to have a basic grasp of how the search works is that it is all based on Z39.50, the protocol used by libraries for sharing information.
The functions and design that are wrapped around the search engine to make it easy to use are the results of several projects. OPLIN spent some time in the ScanPath usability lab at Kent State University observing how people search for information. (We were impressed by the degree to which Google has shaped everyone's expectations of how a search page should behave.) We hired web design professionals to do the graphic design of the page. Finally, we did a lot of in-house programming to manage the ways information is collected from the user and then passed to the search engine, and how the search engine results are subsequently displayed to the user.
You can see the results of all this work by trying a few searches at www.ohioweblibrary.org. While the search is fast and displays results with good relevancy ranking (or ranking by date or title, if you prefer), it is not a replacement for Google. Rather, it is an extra tool for finding information. Google is still the better tool to use if you are looking for a web site. But if you are looking for a document, ohioweblibrary.org performs very well.
OPLIN staff will be presenting information at conferences and workshops over the next few months to give librarians the tips and tricks they need to become Ohio Web Library "power users." Meanwhile, we welcome any and all suggestions for improvement.
UPDATE: Thanks to Matt Weaver of Westlake-Porter Public Library, a screencast about OWL is now available.