The Reader's Helper: A Personalized Document Reading Environment
Paper-based Reader's Helper
As mentioned above, many people prefer paper documents for reading since paper is higher resolution, higher contrast, more portable and readily supports reader note taking and highlighting. To provide for this, a paper-based RH has been created that again adds content to the original document by printing the document in a special way. Figure 5 shows a portion of a document printed by the RH. This figure shows a cover sheet (on the right) and the first page of the document (on the left). The cover sheet contains a complete thumb-nail representation of the document with annotations portrayed as red text. Unlike the electronic version of the Thumbar, the paper version portrays the thumbnail text using a very small typeface so that word presentation and text structure are enhanced. This version of the thumb-nail is also paginated and laid out in columns with pages starting from top to bottom, left to right, each numbered in a small margin to the right of each image. The reader may use the cover sheet to quickly scan the document for relevant areas. At the top of the coversheet is the title and document information . The list of the top three topics of interest found to be relevant in this document are printed below the title . This information is also displayed at the bottom of every page in the document .
Each page in the document contains six thumb-nail images of the surrounding five pages along with the current page in the left-hand margin . This allows the reader to look ahead in the paper document to view the surrounding pages and their relevance with respect to the topics of interest. The annotations in the document are presented as yellow boxes surrounding the key phrases . The page number margin to the right of each thumb-nail image is shaded when the reader is viewing that page. The shaded area moves downward as the reader turns the pages (moves forward in the document) to indicate the page location in the thumb-nail image. On documents longer than 7 pages, the shaded area does not shift downward once the reader turns to page 3. Instead, the thumb-nail image is shifted upward, similarly to the online version of the Thumbar. The shaded area stays in the third cell to show the surrounding context for the current page being viewed in terms of past and future pages in the document. When the document gets closer to the end and the last page of the document is visible in the thumb-nail, the shaded area is again shifted downward until the end of the document is reached.
UNDERLYING TECHNOLOGY
The underlying technology used in the RH is described below. This includes reader profiles and the content recognition method.
Reader Profiles
The RH uses reader profiles to define the topics of interest for each reader. The topics of interest are called concepts since they tend to represent a meaning which the reader would like the system to recognize when analyzing documents. Concepts are defined by a collection of keyword phrases which represent the overall topic of interest. For instance, if a reader is interested in the concept of Intelligent Agents, several keyword phrases can be used to define this topic: adaptive agents, personal assistants, Patti Maes, cooperative information agents. When a document is processed by the RH, a probability is generated for each concept representing the likelihood the document is about that concept. This likelihood is called the concept similarity measure. At present, concept definitions are hand-coded by the reader in the profile. The process is actually quite simple and only involves defining a name for a concept and populating it with the keyword phrases.
