Dissertation/Thesis Abstract

Temporal-informatics of the WWW
by Adar, Eytan, Ph.D., University of Washington, 2009, 213; 3394202
Abstract (Summary)

The World Wide Web is generally experienced as a single snapshot: the Now Web. The Now Web perspective means that only very recent states of the Web are being observed through browsers or through the crawling and index structures of search engines. Such a view ignores the utility of historical data to system designers and end-users alike. This dissertation characterizes content and behavior within the Dynamic Web. This work further illustrates the power of Dynamic Web data by utilizing measures of content change and behavioral measures of use to identify likely targets of user revisitation on the page level. Additionally, two systems are offered for manipulating temporal Web data: (1) DTWExplorer, a tool for analyzing time-series representations of content and behavior to identify correlations in different types of temporal Web data. (2) Zoetrope, an end-user tool for querying, aggregating, and visualizing historical Web data from the context of the Now Web.

This work capitalizes on the richness of the Dynamic Web and demonstrates how moving away from temporally insensitive models and tools can enhance existing applications (e.g., crawlers and search engines) and lead to novel applications and systems.

Indexing (document details)
Advisor: Weld, Daniel S.
Commitee:
School: University of Washington
School Location: United States -- Washington
Source: DAI-B 71/02, Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer science
Keywords: Content change, Dynamic Web, Querying, Temporal Web data
Publication Number: 3394202
ISBN: 9781109609974
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest