Slashdot Mirror
Adobe Building Zoetrope, a Web "Time Machine"
Khuffie writes "Adobe, along with the University of Washington, are developing Zoetrope, an application that will offer a dynamic new view of the web. It is hard to explain on paper, but you can see a brilliant video of the application in action. Essentially, Zoetrope will allow users to travel back in time through a website, and see how the website gets changed. A user can create lenses on the website, for example, focusing on the price of a DVD at Amazon, and see how the price went up and down over the coming months. More interestingly, you can link lenses together across different websites, and for example, see how the price of gas was affected by say, the aggregated google news result of 'war.'"
From the blurb:
More interestingly, you can link lenses together across different websites, and for example, see how the price of gas was affected by say, the aggregated google news result of 'war.'"
Actually, no... You can't use this tool to see how the one thing was affected by the other. You can see how they both changed with respect to time, but that isn't the same.
Please to keep in mind the famous Slashdot Mantra: Correlation is not causation.
The system is limited, however, by how much historical data is available. To test the tool, the researchers chose 1,000 frequently updated websites and stored information captured every hour over four months.
But for Zoetrope to cover the entire Web would mean capturing huge amounts of data, says Eytan Adar, a PhD student at the University of Washington who was involved with the research. He has investigated the rates at which people tend to check different pages for updates and says that such information could provide insights into how often pages need to be recorded, thereby reducing the amount of data that needs to be stored. "It's impossible to crawl and capture some of these things at the rate at which they're changing," Adar says. "But for something like Zoetrope, it's a smaller percentage of the Web that we want to track. We don't actually need to get every single page that's out there."
To make any money, the Zoetrope people will either have to sell this application to websites or setup their own very limited search engine with ads. And if they go search engine style, they'll have no historical data.
It's a neat idea, but the practical applications are still questionable at best.
[Fuck Beta]
o0t!
I think the whole point of this is the analysis capability. It's not just snapshots of old web pages. For that matter it might use archive.org as its data source.