In Search of Better Search

One of the big problems in using primary source documents is finding them. The archives tend to be large making it hard to find the desired resource, and hard to navigate, as they are not typically organized by the criteria needed by student.

There has been a good deal of research in usability issues surrounding search, and the work of UC Berkeley’s Marti Hearst offers some insight on how to make the process easier and more rewarding. Her concept of “faceted metadata” allows users to graphically create complex conjunctive Boolean searches that have been shown to provide easier, faster, and more satisfying search results than standard methods. (Yee, Swearingen, Li & Hearst, 2003).

It works by using metadata (words tags that categorize the resource in some way, such as date, location, etc.) that can be graphically selected to form one or more facets (parts) of a search term. As each facet is changed, only the results that match that facet’s tag are shown.

In our case, we chose to have three facets: A slider pair that that shows only resources occurring between two dates, and radio buttons that limit to display of items that correspond to a person, place, thing, or event. Finally, the students can type in search words to further limit the myriad of choices.

The results are dynamically updated as the buttons or sliders are changed, offering instant feedback to choices. A link to a demo version appears below.

Links

PrimeTime Demonstration
www.digitalschoolbook.org/hist/script.html

The Flamenco Search Interface Project
bailando.sims.berkeley.edu/flamenco.html

Faceted Metadata for Image Search and Browsing
bailando.sims.berkeley.edu/papers/flamenco-chi03.pdf