A View from Emerging Technology from the arXiv
How to Carbon-Date a Web Page
If a Web page lacks a time stamp, how do you know when it was created? A new Web application could help.
Ever needed to know the age of a Web page only to discover that it lacks a time stamp saying when it was published?
If so, then the work of Hany SalahEldeen and Michael Nelson at Old Dominion University in Norfolk, Virginia, may be of interest. These guys have created a Web application called Carbon Date that works out the creation date of a page by searching for the earliest evidence of its existence.
The process is straightforward. Many Web pages end up being recorded in various ways soon after they are created. For example, it’s easy to check Bitly to see the first time that anybody shortened the URL in question, or to use Topsy to check the first date that anybody tweeted the URL. Then there is Memento, which reveals the first time the URL was recorded on a Web archive. Google can also reveal the first time the page was indexed, and even the last-modiﬁed HTTP response header of the page itself shows when it was last changed.
Each of these is straightforward to check by itself, but checking them all to find the earliest date is time-consuming. Carbon Date automates the process.
SalahEldeen and Nelson say their new tool works reasonably well. They tested it on a set of 1200 Web pages for which the creation date was already known. “We were able to estimate a creation date for 75.90 percent of the resources, with 32.78 percent having the correct value,” they say.
That’s not quite as accurate as a researcher or journalist might like, but it’s a start. If you wanto to test it yourself, SalahEldeen and Nelson say they’ve made Carbon Date available at http://cd.cs.odu.edu/cd/<YOUR_URL_HERE>. (However, at the time of writing it did not appear to be working.)
Update 23 April: link updated and now working. See comment below. The service gives the estimated date of creation of www.technologyreview.com as October 2001. The historians at Tech Review tell me they first started using the domain in February 2001, so that’s not far off.
Ref: arxiv.org/abs/1304.5213 :Carbon Dating The Web: Estimating the Age of Web Resources