The Archival Acid Test: Evaluating archive performance on advanced HTML and JavaScript

Abstract

When preserving web pages, archival crawlers sometimes produce a result that varies from what an end-user expects. To quantitatively evaluate the degree to which an archival crawler is capable of comprehensively reproducing a web page from the live web into the archives, the crawlers' capabilities must be evaluated. In this paper, we propose a set of metrics to evaluate the capability of archival crawlers and other preservation tools using the Acid Test concept. For a variety of web preservation tools, we examine previous captures within web archives and note the features that produce incomplete or unexpected results. From there, we design the test to produce a quantitative measure of how well each tool performs its task.