Tuesday, April 26, 2011

Paper Reading #23: Automatic Generation of Research Trails in Web History

Comments:
Adam Friedli
Zack Henkel

Reference Information:
Title: Automatic Generation of Research Trails in Web History
Authors: E.R. Pedersen, K. Gyllstrom, S. Gu, P.J. Hong.
Presentation: IUI'10, February 7-10, 2010, Hong Kong, China

Summary: This conference preceding discusses the development of an interface providing researchers with automatic generated trails of their research. Based on an ethnographic study, they realized the lack of such a system that would provide users with information about the websites they visited, when they visited them and what information they could find in there.

As previous works they describe the History features provided in most web browsers, and the Google History feature. However, these interfaces are not specific to one topic, and do not provide users with more information but URLs. Their system would actually keep track of individual research trail objects, it would analyze the data semantically as wells as the events of each visit.

Their system is currently implemented using a user interface and a model server. Their user interface can be obtained from the New Tab feature that most web browsers have. From there, users can see the most recent research trails. Their model server is currently using Google history data. They assessed their development internally, by using colleagues history data, and as they were developing the system, they were assessing it using this data.

Discussion: This is a great idea! I can't believe there isn't something like this available already. I think it has happened to everyone that while doing web searches we forget what website we were on, and how to get there. Something else I liked was how they describe their motivation, and how they believe this tool can be useful for people performing different types of research, not just continuous research. If someone goes back to work on something they haven't worked on for a while, they can know exactly was was the last things they worked on.

However, I would have loved to see some images of their implementation, and a user study. With a user study they can find out more accurately how appealing this would be to users. We know it would be useful, but would users really use this tool?

Source
The picture above show the current implementation Google provides for its users. Something similar and mentioned in the paper.

2 comments:

  1. To me, this seemed like a tool that could just operate in the background. It might even be able to be built straight into Google. If this ends up being the case, I hope it does not become exclusive to the Chrome browser. I would love to have this as a browser plugin of sorts to experiment with at some point.

    ReplyDelete
  2. I couldn't believe something like this wasn't already available either. It's a pretty cool idea.

    ReplyDelete