Researchers at the University of Glasgow's Department of Computing Science have been working for the past three years on a project that's now called Terrier. Terrier stands for TERabyte RetrIEveR. Terrier is cutting-edge software for the rapid development of Web, intranet and desktop search engines. It is a proejct funded by a UK Engineering and Physical Sciences Research Council (EPSRC) grant. The Terrier project team currently consists of three researchers, five PhD students, and five programmers.

According to the Terrier project website, it is written in Java and has been "successfully used for adhoc retrieval and Web search and cross-language retrieval in a centralised or distributed setting...Terrier originated from a framework initially developed by Gianni Amati. Since then, more people became actively involved in extending and optimising this framework. Terrier has been used for conducting fruitful experimentation, with excellent outcome, allowing for a better understanding of theoretical Information Retrieval."

Some of Terrier's features include hyperlink structure analysis, a combination of evidence approaches, automatic query expansion/re-formulation techniques, query performance predictors, and compression techniques. Most current commercial search engines incorporate a link analysis component in their document ranking algorithm, such as Google's PageRank and IBM's HITS (Teoma.com). Terrier is different because it includes a novel link analysis component that is more general than Google's PageRank, does not use parameters such as the damping factor, can be applied in a query-dependent or independent way, and could be used in various applications, such as multilingual retrieval. Other features include length normalisation, a retrieval approaches selector, and dynamic selection of Web retrieval approaches. It includes a set of performance predictors and has a very low computational overhead.

Terrier is available for download as open source software under the Mozilla Public License (MPL). You can download the open source version or find out more information about it at the project's website (http://ir.dcs.gla.ac.uk/terrier/).
March 14, 2005





Bill Hartzer currently runs a Strategic Online Marketing Consultancy that includes services such as search engine optimization, social media marketing, and online reputation management.

Bill Hartzer formerly managed the Search Engine Marketing division of Vizion Interactive and MarketNet, leading interactive marketing and website design firms in the Dallas, Texas area.

Hartzer is a successful writer, blogger, search engine marketing, and social media marketing expert. During the past fifteen years, some of his many accomplishments include: Search Engine Marketing Manager, Vizion Interactive, Search Engine Marketing Manager, MarketNet, Search Engine Optimization Strategist, Intec Telecom Systems PLC, Webmaster, Intec Telecom Systems PLC, Founder, Dallas/Fort Worth Search Engine Marketing Association, Owner/Author, Corporate Web Site Marketing, Administrator, Search Engine Forums, Frequent Speaker, Search Engine Strategies Conferences, Frequent Speaker, WebmasterWorld's PubCon Search Engine and Internet Marketing Conference.








Search Engine Guide > Bill Hartzer > The Terrier Project