Skip to content.
Sections
Personal tools
You are here: Home » Features » Search Service Projects

Search Service Projects

Neil Fegen
Last modified 24 Apr, 2006
Published 24 Apr, 2006
Neil Fegen, Assistant Co-ordinator of the CETIS Metadata Special Interest Group, provides an overview of the JISC funded search related web service projects and illustrates the potential benefits of this work with useful two scenarios.

Introduction

This article provides an overview of current and completed projects which have implemented the Search and Federated Search services from the e-Learning Framework [1]. The article begins by explaining the purpose of the services, and then outlines two simple scenarios to illustrate the use of search projects, before giving a brief overview of each project.

Search-related specifications

The Search and Federated Search services are part of the Common Services strand of the ELF. The Search service supports the finding of information resources – such as learning objects – from a single search type, such as a library catalogue. The concept of Federated Search is simple: search a number of databases simultaneously with a single interface which it is hoped is more convenient for the user than having to log-in to multiple locations and use several different interfaces. A federated search service may also add value by removing duplicate matching records, filtering and sorting the list of matching records by criteria specified by the user or drawing together related resources.

A number of search protocols allow searches to be performed on remote targets such as repositories. Prominent protocols include: Z39.50, a pre-Web client server protocol for searching and retrieving information from remote computer databases, and widely used in library environments; and SRU/SRW, a search protocol for Internet search queries. The OpenURL standard attempts to advise users on the appropriate version of a resource, which may be a link to an online version or information on a physical resource. The OAI Protocol for Metadata Harvesting allows aggregation of metadata from several sources, which can then be searched by an appropriate search tool (but not by the OAI Protocol itself).

Scenarios

Searching, and the subsequent discovery of materials, can have a real impact on teaching and learning. A well-implemented search toolkit within a VLE can allow a teacher or student to discover a vast number of resources which may be unavailable when using a general Internet search engine, and in addition, a search toolkit may allow the creation and recording of reading lists of useful resources. The benefits of the existence of such a search toolkit, and the drawbacks associated with the absence of one, can be highlighted in the scenarios below.

Scenario 1

Barry is a lecturer at Badsearch University, where a lightweight VLE is seldom-used by staff and students. With no dedicated e-learning team, VLE content is populated by lecturers depositing lecture material – typically in the form of documents or images – and students may use the VLE briefly to access these files (which are usually duplicates of notes given out in class).

Barry decides he wants to provide some background reading for his students, and makes a list of some relevant reading resources. The list has been compiled using word processing software, with entries added from various search results, including popular search engines. This process is far from ideal, because it involves either manually typing in the information, or copying the text and pasting it into the document. Furthermore, when looking at useful resources collected previously, Barry cannot recall which search terms were used to find these results and therefore has a static list.

A VLE that allows the saving of search results would have solved this problem, and the ability to create reading lists from the search results would have enabled easily-generated and manageable collections of resources. It would also have been possible to automatically provide – in some cases – links to the appropriate version of the resource, which would have appeared within the search results.

Barry posts a word-processed document into the VLE, which requires students to manually search for each resource. Due to other commitments, Barry did not have time to check for the availability of physical resources, and the list contains references that are not held by the university library. This tedious process causes students to stop after two or three entries, and the list is largely unused, due to the effort involved in trying to find resources.

Scenario 2

Gary studies at the University of Goodsearch, where a dedicated e-learning team ensure a smoothly-run VLE is available for staff and students. The team recently implemented a federated search toolkit, which enables users to search a wide range of targets for content, from just one interface. The results are grouped together, and where possible, links to full-text resources are provided.

The toolkit provides Gary with many search locations, available from one interface. Rather than visiting many search engines, Gary types his query into the text box on the VLE’s search page. The query is then sent to each target that the search tool has a licence to access – with targets including library catalogues, preprint repositories and collections of online journals – and the list of results then presented in an easily-readable format for Gary to browse through. With the implementation of OpenURL in some institutional targets, it is possible for Gary to access the appropriate version of a resource without leaving the VLE. The style of search also allows Gary to search by a number of filters. This is particularly handy when he is searching about a certain lecturer: he can distinguish whether the lecturer is the author of a resource, by including him in the ‘author’ field, or simply mentioned as a ‘keyword’.

By covering a number of targets, Gary finds a large volume of information, and spends some time to filter what is relevant to his task. By creating a resource list – a feature of the toolkit – he can keep track of which resources he has found to be useful, which will not only assist him in compiling his list of references, but also others who want to look for material on similar topics. Gary can also recall this list to revisit resources as required.

The features of the federated search toolkit assist Gary in compiling his deliverable and he achieves a high score for his dissertation.

The projects

JISC have funded a number of projects looking at a range of search-related toolkits and services, which are listed below, chronologically.

RDN Subject Portals Project (SPP)

The RDN Subject Portals Project (SPP) developed portal functionality for five of the eight subject hubs of the Resource Discovery Network, with an aim to provide access to high-quality information resources made easier for the user through aggregated cross searching, streamlined account management, user profiling and the provision of additional services. The project, which was managed by UKOLN, completed its second phase in August 2004 [2].

Middleware for Distributed Cognition (MDC) project

The MDC toolkit project developed a means for academics and students to search a range of reference catalogues, and then build a reading list that could be printed, uploaded to a web site or stored online. Developed by the University of Oxford and University of Edinburgh, MDC searches across different types of databases – such as SRW and Z39.50 – and queries different types of metadata, such as Dublin Core and UK LOM Core. The software is intended to be seamlessly embeddable into learning environments and was finished in October 2004 [3].

D+ (Brokerage for Deep and Distributed e-learning Resources) project

The d+ project developed a search toolkit to aid the discovery of resources in distributed and heterogeneous Z39.50-compliant search targets. d+ was a collaborative project led by the University of Edinburgh and facilitated the underlying search service of the DEVIL utilities (a previous JISC-funded project). d+ was completed in November 2004 [4].

ELF Search Service Demonstrator Project

The ELF Search Service Demonstrator project evaluated the suitability of the d+, MDC and SPP resource discovery toolkits for enabling resource discovery from within a VLE. The usefulness of embedding resource discovery into a VLE was considered, based on the opinions of teaching and support staff, by looking at the Moodle and WebCT VLEs. The project compared each toolkit on a number of criteria, and asked stakeholders to ascertain whether an embedded resource discovery tool (within a VLE) was preferable, along with opinions on enhanced functionality. The project was completed in September 2005 by the Institute of Computer Based Learning at Heriot-Watt University, Edinburgh [5].

Resource List Toolkit

The Resource List Toolkit project developed a software development toolkit (SDK) to mediate the reuse of resource lists. These lists would be created and stored in distributed and heterogeneous systems, such as e-learning tools, VLEs, portals, dedicated resource list systems and repositories. The project, led by the University of Edinburgh, was completed in October 2005 [6].

DICE (Resource Discovery Tools Evaluation and Integration) project

The DICE Project evaluated the d+ (Discovery plus) resource discovery toolkit developed at Edinburgh University Library, comparing it with other resource discovery services, and in integrating such discovery services with their in-house VLE (COSE). The name of the project was derived from d+ and previous work on integrating web based resources (ICE - Integrating COSE and E-resources). DICE was completed in December 2005 [7].

ASK (Accessing and Storing Knowledge) project

The ASK Project will enable users to perform a federated search over repositories which implement the Z39.50, SRU and SRW search protocols. Specifically, the list of repositories will include DSpace, ePrints and Fedora but the user should be able to customise the search by re-defining which targets are used. The project aims to deal with a range of resource types, including images, documents, reading lists and IMS Content Packages. ASK is being coordinated by Oxford University Computing Services, and is due to be completed in May 2007 [8].

PERX (Pilot Engineering Repository Xsearch) Project

PERX is a collaboration led by Heriot-Watt University to develop a pilot service which will provide subject resource discovery across a series of repositories of interest to the engineering learning and research communities. Issues to be investigated include the range and availability of digital repository sources, exploration of cultural barriers, and improving search and browse results presentation. PERX is due to be completed in May 2007 [9].

Conclusions

JISC has funded (and continues to fund) a number of search- and federated search-related projects which cover a wide range of services required for cross-searching – please let me know at nf16@macs.hw.ac.uk if there are any relevant projects missing. The SIG discussion website [10] will be updated to show projects incorporating e-Framework [11] standards.

References

[1] e-Learning Framework

[2] Subject Portals Project

[3] MDC project

[4] d+ project

[5] ELF Search Service Demonstrator Project

[6] Resource List Toolkit

[7] DICE project

[8] ASK project

[9] PERX Project

[10] CETIS Metadata and Digital Repository Special Interest Group

[11] e-Framework

 

Supported by JISC Supported by CETIS
Powered by Plone