Aquabrowser

Aquabrowser was selected for purchase by the University of Edinburgh and was implemented between April and September 2008, to be available for students at the start of our academic year. This was after a limited market review where we identified available vertical search products, established our evaluation criteria and had supplier visits to do on-site demos. The products were assessed on value for money, perceived ease of implementation, technical requirements and functionality. We selected Aquabrowser because we identified it as a low risk, (in terms of cost and staff resource), highly functional tool that would allow us to experiment with users’ reaction to a vertical search product. Aquabrowser was intended to replace the existing interface to our library catalogue which was felt to be dated and not intuitive for new users.

Aquabrowser has a simple “Google-like” single search box and once a search has been done it presents a screen laid out in three parts (see Figure 1). On the left is the word cloud, a set of terms derived from the search results; in the middle are the results themselves; and finally there is a range of limits, or facets, that can be used to refine the search and make it more targeted to user requirements. The word cloud encourages the user to navigate through the results, always focusing their search to find exactly what they want. The word cloud highlights words that have already been used in the search so the user has a trail through their search process. The sophistication of the limits available to us in the system ensures that the user finds what they are looking for. In a way, the word cloud presents a form of serendipitous browsing that has not been possible to represent in previous versions of the library catalogue.

Aquabrowser does not work dynamically on the data held in our library catalogue. This data has to be extracted to be loaded into Aquabrowser. This may seem like additional work but, apart from the initial extract, the daily extracts are done automatically using scripts. A major advantage of this approach is that it allows you to consider your catalogue records as data, outside any constraints imposed previously by your library management system in terms of storage and display. We took the opportunity to completely rethink our approach to our MARC mapping, field labels and display fields as part of this implementation. We feel that Aquabrowser allows us to exploit more successfully the investment we have made in the creation of our high quality MARC records.

As a result of our short timescale much of the initial customization was done by the supplier based on our detailed specification, and then further developed in a focused week of development where the supplier came on site. We dealt primarily with the software developers, Medialab, but our ongoing support relationship is with Infor.

Aquabrowser is shipped with a user administrative control panel which allows the library to customize the majority of the features of the system. Most configuration files are held as XML so knowledge of this is needed to customize the system. Customization can either be done by the customer or by Medialab, as required. Our experience of working with Medialab has been very good; they are responsive and quick to implement changes we have requested.

Aquabrowser comes with a detailed statistics tool which provides data not just on usage but also on favourite search terms, searches that produced few results, how often users used specific limits to narrow their search, how deeply the user went into the results of the search, and how many pages of results they went through, for example. (See Figure 2.) This is much more detailed information than we have had previously and we are excited at the opportunities this gives us to understand how users are using the catalogue and therefore improve how we develop that service.

Morag Watson, University of Edinburgh
http://aquabrowser.lib.ed.ac.uk