The FACET web interface - Demonstrations
The FACET web interface builds on the results of previous project work, to create a dynamically generated, data driven "query builder" web interface that demonstrates many of the practical thesaurus integration ideas that evolved during the course of the project. The interface illustrates how thesaurus data (and particularly term expansion across thesaurus relationships) may be presented and utilised in an intuitive and productive way. These demonstrations do not rely on prebuilt HTML pages; most thesaurus content presented is generated dynamically via the FACET architecture's in-memory representation, the reminder from a database.
The aim is to explore how a thesaurus can be integrated into the search interface and the potential of semantic expansion in querying collections indexed with controlled metadata. The current interface should be considered an "advanced" search interface. We are currently working on simpler web interfaces for searching where more of the term expansion functionality takes place behind the scenes. See the overview for background on the demonstrator and the FACET project objectives.
The user interface
The various features of the user interface are described in the following links, and we would recommend that you take some time to experiment with the 'live' components using the demonstration examples provided, prior to using the full query builder. Please bear in mind that this is an experimental research prototype, not a robust commercial application. Also please note that due to the implementation method chosen (see Technical platform issues below) this interface will only function correctly using Microsoft Internet Explorer browser (Version 4 or later).
The thesaurus used in all examples is the Getty Art & Architecture Thesaurus (AAT) - a controlled vocabulary popular within the cultural heritage sector. Note that as a demonstration we are NOT displaying the most recent version of this vocabulary, and users are directed to the Getty web site for the most up to date information.
The query builder operates over an extracted subset of the National Museum of Science & Industry (NMSI) collections database. Please note that this extract was provided to us for use within the project some time ago, and we have changed the data for demonstration purposes and experimented with some indexing. Therefore any displayed data should not be regarded as current, and does not reflect the cataloguing and indexing employed within NMSI itself.
Technical platform issues
The move to a browser based web environment presents some specific new technical challenges to traditional application implementation techniques - chiefly (1) state persistence, and (2) bandwidth limitations.
The eventual platform adopted was Active Server Pages (ASP). Server-side scripting and compiled server-side components are used for database access, and cascading style sheets are used for presentation.
The interface consists of a set of discrete functional scripting components (implemented as "scriptlets") for dynamic functionality. Screen flicker is eliminated during operation, as each component updates its own data from the server without causing an entire page refresh.
The advantage of this approach is that it allows an interactive web application to maintain state and persist data, even though it is operating within a stateless environment. The disadvantage is that the use of Microsoft's XMLHTTP for client-server communication means this interface will currently only function correctly using Microsoft Internet Explorer browser (Version 4 or later).
Additional note - > *If problems running in IE7* The FACET demonstrators employs (preAJAX) partial screen refresh 'scriplet' technology, specific to Internet Explorer and the demo was developed for IE6. A possible problem with running the FACET web demonstrator on IE7 installations is that IE7 may disable Dynamic HTML (DHTML) scriptlets by default when first installed. They can be re-enabled by the following steps:
We will continue to build on the experiences and outcomes of this project. In the longer term we are exploring the possibilities offered by emerging technologies such as web services. It is envisaged that this will facilitate developing a set of semantic terminology service API's, enabling the dissemination of semantic expansion and associated functionality to a wider audience.
[back to top]
Copyright ©2002 - 2017 University of Glamorgan, all rights reserved