Date of Award

Summer 2011

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

Committee Director

Michael L. Nelson

Committee Director

Yaohang Li

Committee Member

Michele C. Weigle

Committee Member

Mohammad Zubair

Committee Member

Robert Sanderson

Committee Member

Herbert Van de Sompel

Abstract

Given the dynamic nature of the World Wide Web, missing web pages, or "404 Page not Found" responses, are part of our web browsing experience. It is our intuition that information on the web is rarely completely lost, it is just missing. In whole or in part, content often moves from one URI to another and hence it just needs to be (re-)discovered. We evaluate several methods for a \justin- time" approach to web page preservation. We investigate the suitability of lexical signatures and web page titles to rediscover missing content. It is understood that web pages change over time which implies that the performance of these two methods depends on the age of the content. We therefore conduct a temporal study of the decay of lexical signatures and titles and estimate their half-life. We further propose the use of tags that users have created to annotate pages as well as the most salient terms derived from a page's link neighborhood. We utilize the Memento framework to discover previous versions of web pages and to execute the above methods. We provide a work ow including a set of parameters that is most promising for the (re-)discovery of missing web pages. We introduce Synchronicity, a web browser add-on that implements this work ow. It works while the user is browsing and detects the occurrence of 404 errors automatically. When activated by the user Synchronicity offers a total of six methods to either rediscover the missing page at its new URI or discover an alternative page that satisfies the user's information need. Synchronicity depends on user interaction which enables it to provide results in real time.

ISBN

9781124990699

ORCID

0000-0003-0130-2097

Share

COinS