Time-Travel for the Web Comes from the 'Memento' Solution
A team of researchers including Old Dominion University computer scientist Michael L. Nelson has taken aim at the ephemeral nature of Web pages, and the results so far are garnering attention worldwide.
"Memento: Time Travel for the Web" is the title of a new paper and presentation that the researchers have been rolling out this month.
Most Web users understand that a URI (uniform resource identifier) that they bookmark one month may very well return a quite different page a few months later; the original page will have been supplanted by a later version. These same users also probably know that current avenues for retrieving old-archived-versions of Web pages are limited and can be convoluted and slow.
"The Web has a terrible memory," is how Nelson and his colleagues describe the problem in the paper they published early in November about their Memento system.
To enhance this "memory," the researchers have devised a solution that works via software for servers and browsers to allow the browsers to enter "time-travel" mode. This enables searches for a past version of a file that is date-and-time specific, rather than the most up-to-date version. Memento, in other words, can provide an easy way to dig up a news service's Web page from late summer 2005 when Katrina and her flooding were ravaging New Orleans.
The Memento project is supported by the Library of Congress under the National Digital Information Infrastructure and Preservation Program (NDIIPP). Leading the research team together with Nelson is Herbert Van de Sompel, a computer scientist at the U.S. Department of Energy's Los Alamos National Laboratory (LANL).
Other members of the team are Lyudmila Balakireva, Robert Sanderson and Harihar Shankar of LANL and Scott Ainsworth, a graduate student at ODU.
Nelson and Van de Sompel presented the Memento technology at a NDIIPP seminar at the Library of Congress in Washington Nov. 16, and Van de Sompel gave another presentation of the work for the Online Computer Library Center's Distinguished Seminar Series in Dublin, Ohio, on Nov. 19.
An article posted on the New Scientist Web site in mid-November gave Memento a boost, Nelson said. It spawned similar articles in media in the United States and as far away as China. A posting about Memento also appeared Nov. 18 on The Chronicle of High Education blog "The Wired Campus."
Memento works via a function of the hypertext transfer protocol (HTTP) that supports the World Wide Web. HTTP defines how Web pages are formatted and transmitted from servers to browsers.
A function called "content negotiation" allows latitude in how the browser interacts with the URI server. For example, the page request settings of the browser that contacts the URI might dictate that the page be sent in French as opposed to English. The settings might also show a preference for HTML over PDF, or for GNU Zip files over Zip files.
Nelson, Van de Sompel and their colleagues have designed Memento to use a new dimension of this page request function that negotiates for a specific date and time.
On a server running the Apache Web system, four lines of extra code will build in the date-and-time negotiation, according to the researchers. On the browser, a drop-down menu will let users enter the date and time for the page they want to view.
The Memento research paper notes that most archival copies have URIs that are not protocol-connected to the URI of the original resource. "This turns accessing archived resources into a significant discovery challenge for both human and software agents, which typically involves following a multitude of links from the original to the archival resource, or of searching archives for the original URI," the authors state.
"This paper proposes the protocol-based Memento solution to address this problem, and describes a proof-of-concept experiment that includes major servers of archival content, including Wikipedia and the Internet Archive. The Memento solution is based on existing HTTP capabilities applied in a novel way to add the temporal dimension. The result is a framework in which archived resources can seamlessly be reached via the URI of their original."
As Nelson sees it, "The Memento project advocates a rather straightforward approach to make navigating last year's Web as easy as navigating today's. Remnants of the past Web are available, and there are many efforts ongoing to archive even more. It's just that the past Web is not as readily accessible as today's.
"For example," Nelson added, "if you want to see an archived version of http://cnn.com, you can go to the Internet Archive's Wayback Machine and search for it there. But wouldn't it be much easier if you could just connect to cnn.com indicating that you are interested in the pages of, say March 20, 2008, not the current ones. You could activate a time machine in your browser or bot!"
More about the Memento project, including a "demo" link to an experimental Memento client, can be found at http://www.mementoweb.org. A FireFox "add-on" will be available shortly to enable Memento in a user's personal web browser. There is also project information on the blog of Nelson's Web Science and Digital Library Research Group at http://ws-dl.blogspot.com.
Nelson, an associate professor of computer science at ODU who won a $500,000 National Science Foundation Young Career Development Award in 2007 to pursue his innovative ideas about digital preservation, said the Memento work is far from over. Several bugs need to be ironed out, among other things to avoid time-travel Web experiences that turn up previous versions, but not the versions that were sought.
Also, for Memento to catch on, Website owners must store many more time-stamped versions of their pages than they do now. Nelson said he hopes the introduction of Memento will encourage them to do so.
Van de Sompel told New Scientist, "I would love to see Memento supported. It would be such fun to set our browsers back in time and just browse the past."