TopBlend

Structured diff:

TopBlend: An Efficient Implementation of HtmlDiff in Java

ABSTRACT

The World Wide Web is growing rapidly with new and changing web content. Detecting changes in web pages is crucial for website masters who care about website integrity. It is also convenient for web surfers who are constantly looking for new products, services, or information on their favorite websites. This paper describes TopBlend, a new HTML differencing tool implemented in Java. TopBlend uses the fast Jacobson-Vo algorithm, which solves the Heaviest Common Subsequence problem, for page comparison. Performance results indicate that TopBlend significantly outperforms a previous HTML differencing tool in most time-consuming jobs, often by 1-2 orders of magnitude. TopBlend allows comparisons to be performed either on the server or client side. The latter has become increasingly powerful enough to offload busy servers by performing heavy computations. TopBlend can present the results in either a merged HTML view or a more convenient side-by-side view for web pages with complex graphics designs. TopBlend has been integrated with the AT&T Internet Difference Engine (AIDE) and can also be employed easily by other website tracking services.