Imagine Google without a search box

If you are looking for audiovisual content it is hard to know where to start. Google can search vast amounts but there is a whole section on the web that is only available to education, and Google by itself is not enough for scholarly use.

When we started creating a search environment to look at multiple databases in one we were given the oddest, but most precious piece of advice: ‘lose the search box’. The stunned looks around the project team’s faces said it all… isn’t searching about, well, searching?

At the BUFVC we offer access to many different types of datasets, from listings of practically everything that was broadcast on British TV and radio since 2001, to newsreels produced throughout the 20th century. In the United Kingdom there is no audiovisual equivalent for the British Library legal deposit. We therefore work hard to fill that information gap and find ways for education institutions to gain access to audiovisual content.

When we thought of ‘federating’ most of our data (searching it all at once) we thought it would be easy. And it was. We know our data well and could map it quickly.  But that wasn’t the real challenge. It was by looking at how users interacted with the developing system that we learnt the key lessons. If you are searching our 13 million records there are two difficult steps: what to type in that search box and how to then get rid of all the stuff you don’t want.

The answer to the first dilemma – how to stop users worrying about the empty search box – was to make it less important. We needed to make sure that whatever the exact search term, users should be able to make interesting discoveries even if the results were not the best. To do this we created relationships between our records based on semantics. This means that users will see suggestions for searches and records that may take them in different (but related) journeys.

The second dilemma is all about filtering. The search results page became our control centre, allowing users to filter results, tweak searches, see the suggestions for related searches and use a variety of additional tools. All this makes for a busy page but after many rounds of user testing we think our designer hit the right balance between complexity and elegance. Some of these features include:
•    Human-friendly filters such as identifying results by availability (‘can I see it online, do I need to order it from someone?), media type (moving image, audio, documents), genre or collection
•    A comprehensive history function that keeps track of viewed records, searches and tweaks to searches and the ability to mark and cite or export records in a variety of standards

When reflecting back on the project, I remember mostly the sense of fun, the permission we gave ourselves to think creatively, and our exciting user testing rounds. Developers and users don’t often mix but with good moderation these sessions proved magical. These are three elements I would urge anyone to replicate in future projects.

We are not alone in this development; sites that enable discovery and aggregate results from multiple collections are increasing in number – Edina’s Jisc Media Hub is well worth a visit. We hope to have created something which rewards users with new connections through discovery. Work will only be completed in September and there are many plans beyond that. In the meantime if you are looking for moving image and sound, be it television or radio, newsreels or commercially available programmes for education pay a visit to our beta site and give us your feedback. If you are reading this from a UK higher or further education institution, please remember to login to access all areas.

The open-source based software and interface will be released as a package under an Open Source licence later this year. This project was funded by the Arts and Humanities Research Council and led by Royal Holloway, University of London.

Luis Carrasqueiro is Chief Executive, British Universities Film & Video Council
luis@bufvc.ac.uk