Adrienne LaFrance writes in the Atlantic in an article called “Raiders of the Lost Web” about vanishing websites and the recovery of one particularly important investigative reporting piece, “The Crossing.” Along the way, LaFrance discusses the state of web archiving through brief interviews with Jason Scott and Brewster Kahle of Internet Archive as well as Alexander Rose of the Long Now Foundation. The problems of preserving everything are still insurmountable. The problems of just preserving and serving valuable websites, those curated say or deemed to be of ongoing importance, is possible but not yet done well.
I’m glad to see the continued attention to this problem. With the help of support from IBM research grant entitled “Keeping the Historic Web Alive and Accessible: preserving and serving,” I’m involved in using virtualization to keep older versions of older software open and accessible.
Despite the great efforts of Internet Archive and others, archivists attempts at preserving websites have largely been a collection of web page collected periodically by such successful software as Archive.org’s Heritrix. While page collectors provide a good service especially for sites that do not allow access to the database that drive them, they fail on several fronts.
First, a complex site can never be completely captured by a screen capture program or even a video. The number of potential pages and interactions cannot be duplicated in such a manner.
Second, user interactions with a complex site cannot be duplicated by a captured set of static pages. Understanding interactions is important to the use of archival material.
Third, the context of use and interactions is lost when the user experience is changed from one of interactivity and customization to one of simply retrieving selected static pages.
Even worse are archival activities that attempt to preserve by atomizing the objects from the pages removing images, audio, texts, etc and placing them in new decontextualized object databases completely unrelated to their original context.
We propose, using ibiblio as an example, to attack this preservation problem by using machine virtualization and virtual machine management tools. One suite of tools that particularly interest us are NCSU’s Virtual Computing Lab known as VCL. VCL is in use at over 200 institutions and has a strong support structure. Open source and licensed under the Apache project licenses, VCL is mostly used to provide temporary virtual machine for training computer scientists. We and our partners at NC State see a new use for VCL as a way to manage and present older servers Running older software in a secure yet living and interactive way.
I hope that we and those others of us working on digital preservation and in virtualization can provide a framework that benefits the creators, the archivist, and publics present and future.
Look for some initial results to be posted here soon.