http://www.cs.dal.ca/~jamie/grad/textUI.html
&
.
Created on 21 December 2001 by J. Blustein.
Last updated on 03 December 2003 by J. Blustein.
I want to develop (and evaluate) better browsers for reading on the Web. I want to see how certain features could improve users' experiences:
Much of this can be supported using Java code from the multivalent documents project at UC Berkeley and XML or XHTML.
I'd like this work to grow to be something that can help people to read journal articles and textbooks using portable tablet-style computers or other paper-like interfaces.
This is a complex, long-term research project. Some of the parts can be done as Masters projects and theses. Other parts can be done as PhD work. If you are interested in this research, or any part of it then please come visit me in office 223. A description of some of the parts is below.
I will pay assistantships for qualified students.
The rest of this webpage is organized like this list:
What goals are people trying to achieve when they read journal articles and textbooks? How do they try to achieve those goals?
A task-analysis is a detailed study of users' goals and the sequences of actions they do and objects they use to achieve those goals. Task analyses are independent of the technology that supports those tasks.
We are particuarly interested in three aspects of using text, namely using a glossary, annotating (notetaking), and searching. Those three areas will need analysis. There could be a thesis in each of these or perhaps the analysis will need to be combined with some design.
We are particularly interested in how and why people make marks (writing or other types) on paper copies of journal articles and in textbooks.
How do people deal with unfamiliar words and uses of words in technical writing? When and how do people seek definitions of words? Do readers keep their own list of words? What type of computerized interface is likely to help people? (Early studies suggested that if the interface was not suitable then readers would not use it, even though they could not understand the text without it.)
Why do people search for things in articles? How could we help them to find what they are looking for? Searching includes text and non-text.
What are essential characteristics of readers of journal articles and textbooks? We know, from many studies, that success or failure of people using electronic texts depends, at least in part, on the psychological profile of the user. But we don't know exactly which factors matter.
The user analysis will describe all of the users of the system and how the system could fit into their lives. It will need to answer questions such as
What type of interface will best support users of electronic journals and electronic textbooks? This encompasses the software architecture, the data organization (probably in XML), and the interface.
Although we don't have a design or architecture yet, it is clear that certain tools will need to be built and that interfaces for them will need to be constructed too. We'll need the following tools: a glossary, bookmarks, support for annotation, and automatic dynamic tables of contents.
A list of technical terms and their definitions. The list can be for a class of texts (e.g. computer networks) or one article. There are some published automatic glossary tools.
A Masters student will develop a glossary that users can update themselves (for a single article or a class of texts), and to implement a published algorithm for an automatic system too. The interface component of this work will be small. We'd like to test the tools with users too.
Users will need to be able to tag sections of the text with the equivalent of post-it note flags. The user must be able to select a point in the text, or a region of the text, and assign a label to it. These labels must be something that users can jump back to later. The physical label markers should have colour and text associated with them. It would be best if the labels could be assigned to classes by colour (example of classes might be: important, quote this later, paused here).
Making notes on a text is a major reader activity. Electronic text that does not support this activity and support it well will not succeed. Two types of annotation are considered: plain text annotation and annotation including hypertext links. Two interfaces are considered (although others are possible): notes in the margins of documents and floating notes (like post-it notes on paper, and electronic annotation in Adobe Acrobat).
There is a Masters project in developing an XML schema for representing linked annotation and implementing some tool to do it.
There is at least one project in implementing the two interfaces. It may be harder but the use of the Multivalent Document toolkit and some XML-based tools should make it easier.
Fisheye views present local context in greater detail than distance context. The SuperBook project from Bell Labs implemented a dynamic fisheye view table of contents (using Lisp). That tool should also be incorporated into a complete reading system.
The algorithm is very simple so this can be combined with some other Masters project.
The interface and presentation will need to be worked out later, but as a preliminary idea something like this would work: a floating box in the upper right corner of the screen with the current section number and title. The contents of the box will be updated automatically as the user moves from section to section. This will mimic the running heads in books and journals. There will need to be a way to change the section title to the fisheye table of contents and there may be a need to have the table stay open in some other part of the window too.
Schemata will need to be developed for the following:
This could be a simple project or it may be complicated. If you are interested in it then please discuss it with me.
The tools will be useless without a good interface for people to find and use them. It is hard to know what type of interface we will need but there seem to be two that are worth exploring. Both can be represented on a regular monitor, but a tablet computer would be better.
Developing each type of interface using tools already constructed will be a separate Masters project.
The interface will look like a piece of paper on a desk with a container of tools (bookmarks, post-its, etc.) nearby. The glossary will appear as a small floating window.
The text will be presented in a window with large margins. Link descriptions will be in the margins beside the link text. Post-it type notes can be attached to the document at any point. Bookmarks, etc. will be available from a floating toolbar.
This interface will look like a web browser. The window will be divided into two unequal frames, the larger frame on top. The main text will be in the larger frame and will have wide margins with link descriptions beside them. Ancillary materials such as footnotes, glossary entries, references to cited articles, will appear in the bottom window. Most important: the start of link destinations will appear in the lower window. The contents of the lower window will be updated when the mouse rolls over a link, clicking on the link will update the larger window (as web browsers do now). The tools will be accessible from the usual type of pull-down menu.
There is more to this research than is described on this webpage. If you are interested in this research, or any part of it then please come visit me in office 223.
See also the list of all my suggested thesis topics and areas of interest.
&
.
Created on 21 December 2001 by J. Blustein.
Last updated on 03 December 2003 by J. Blustein.