Tuesday, April 19, 2011

Paper Reading # 22

Comments
Comment1
Comment2

Reference
DocuBrowse: Faceted Searching, Browsing, and Recommendations in an Enterprise Context
Andreas Girgensohn, Frank Shipman, Francine Chen, Lynn Wilcox
IUI'10, February 7-10, 2010, Hong Kong, China


Summary
This paper talks about browsing and searching for documents in enterprise document repositories and the system the author propose called DocuBrowse. Their main focus is searching for information in a unstructured corporate document repository. The weaknesses of current enterprise document repositories are that the employees usually know where the document they are searching for is located. In a corporate document structure, information is created in such a way that documents are stored in files corresponding to projects or a certain department within the corporation. Finding this documents is easy when the user expects the document to be there. The authors propose a system where document searching resembles more of a web search engine where you can search for documents that you are unfamiliar with, there does not need to be a repository structure and most importantly a user does not need to expect a specific result.  Browsing in DocuBrowse is similar to the famous directory trees. In order to propose an innovative interface design, the authors use a data-oriented document analysis similar to the search engines (Google). DocuBrowse supports faceted searching and also adds the capability of recommendations. This way users will be allowed to determine if their search was successful or if some properties of the document can lead them to the correct information.


Discussion
While working on my internship I became really frustrated sometimes when I was trying to search for information that I did not know where to search for it. The system we were using suffered from the same weaknesses this paper points out and I truly believe DocuBrowse could have been really helpful. It is no lie that current web search engines are extremely powerful and it is an interesting idea to add those capabilities to enterprise document repositories.

No comments:

Post a Comment