Visualizing the World Wide Web: Putting the User in the Driver's Seat

Organized and chaired by Nahum D. Gershon, The MITRE Corporation

Summary Description

Information is dispersed over many Internet resources and quite frequently, users feel lost, confused, and overwhelmed. The panel and the audience will discuss how advances in interactive computer graphics and visualization software and hardware could make the information distributed over the Internet more intuitively searchable, accessible, and easier to use by people from all walks of life and interests. This will enable the us to make full use of the Internet's information universe from our computers.


Extended Description

We live in an exciting time. Connecting numerous information-stuffed computers dispersed around the world has created an exciting universe of information. This information revolution has enabled us to explore this universe from our computers. However captivating, we still have a long way before the use of this information universe is easy and intuitive.

Information is dispersed in many sources over the Internet and at times users feel lost, confused, and overwhelmed (justifiably so). To find required information or to browse through information, users need nowadays to confront frustrating searches through arrays of user- debilitating menus and belligerent computer systems. Some of the remote sources are massive and once the user has got the information, he or she needs to browse through large amounts of text, data tables, and images. How should the user know where the sources of the relevant information reside, how to get them, and, once the sources are retrieved, how to get the relevant information from them?

The World Wide Web (WWW), developed at CERN, Switzerland, and Hyper-G, developed at Graz Univ. of Technology, allow the user through a set of menus to roam through information spaces of documents or images. Captivating browsers, such as NCSA Mosaic and also Harmony have transformed the process of getting information off these Internet distributed information systems. However, some major difficulties still remain.

Advances in interactive computer graphics hardware and mass storage have created new possibilities for information navigation, retrieval, and access in which visualization and user interface (UI) could play a central role. The question is how to utilize the advances in graphics technology and experience to reduce these frustrations and lower the time and cost of navigating through the information dispersed over the Internet, finding specific information, and accessing it once found.

If we do not involve the users in designing the information highway and its interfaces, we will create useless information systems. As long as there is a human being sitting in the front of the screen, the user interface of the information highway needs to be user-oriented (UO), taking the user needs into account. To be able to do so, we need to understand

These are some of the difficulties and issues confronting the design of human interfaces to the Internet and other distributed information systems of the future. And, there are more:

Panel Statements

Putting the User in Charge of the Information Space Using Visual Skills-- Nahum Gershon

One of the major problems of current information systems distributed over the Internet is that the information is rigidly put in place. Pieces of information are linked together in a rigid structure-- no changes are allowed. However, these pieces of information could be related to each other in various ways depending on the application, problem, personal way of thinking and perception, or culture. For distributed information systems to be effective, they should allow each user to construct his or her own information space with links and associations (among pieces of information and whole documents and images ) that fit the problem, application, or ways of thinking and perception.

Another major problem facing systems of today is that while surfing over the Internet, users often do not know where they are in information space and do not remember how they got there. In short, users are lost. One solution is to provide users with both a local and a global view of the information space. These views should be represented visually to promote quick perception and understanding. The user can "jump" from one document to another by clicking the mouse button without the necessity to go back resource by resource. This eliminates the necessity to go back "page by page.

Enabling the user to modify interactively the links among the documents and images using a visual display and to (visually) view the information space globally and locally have been implemented over the World Wide Web in a MITRE enhancement to NCSA Mosaic.

Visualizing Global Networks -- Stephen G. Eick

The Internet's accessibility, richness, growth, and popularity makes it ideal for exercising network visualization tools. Effective, task-oriented visualizations will make features of the Internet immediately recognizable by highlighting patterns and enabling the user to discover interesting and useful insights about the Internet. The Internet, as with all networks, consists of nodes, links, and statistics associated with the nodes and links. The statistics may be time varying, as may the network structure. The link statistics may involve packet traffic between the sites or the number of users following specific site-to-site hypertext web links; the node statistics may involve the number of accesses to particular home pages.

Research focus on network visualization needs to involve inventing the visual metaphors, interaction techniques, and infrastructure for displaying large networks. Methods to display networks often involve node and link diagrams, with the glyphs encoding the node statistics positioned spatially, perhaps in 3D, and lines drawn between the nodes encoding the link statistic. The usual methods for creating node and link diagrams work well for small networks, but for large networks, such as the Internet, the displays become cluttered and visually confusing. We have developed three network visualization systems that solve the display clutter problems in different ways using layout algorithms, 3D, and animations to show patterns andcommunities of interest. These techniques allow for visualizing networks with hundreds of thousands of nodes and links, and statistics many time periods.

Navigation on the Web-- Joseph Hardin

In the last 24 months, hundreds of thousands of people have been introduced to hypermedia through World Wide Web browsers like NCSA Mosaic. These tools provide users with the ability to easily browse through the global online information mass by simply clicking on hyperlinks. An early addition to the textual hyperlink was the ability to map links to images or portions of images. Anything that could be placed in an image with a format like GIF could be made into a link or set of links. This resulted in a rapid blossoming of methods on the Web to provide people with visual signposts to online resources. Maps, pictures of buildings, images of floor plans, aerial views of cities all were put up on the Web and used as guides to a variety of subject domains. This portion of the panel will tour some of these examples of visual navigation on the Web, discussing the ways authors have utilized graphic navigation, and the advantages and limitations of current practices.

Browsing and Navigation Through Deep Hyperspace-- Frank Kappe

One of the big issues in finding information in the Internet is what is known as the "lost in hyperspace" syndrome: users cannot get an overview, cannot find specific information, stumble over the same information again and again, cannot identify new and outdated information, cannot find out how much information there is on a given topic and how much of it has been seen, etc.

I can see three counter measures to deal with this problem: reasonable a- priori organization of information, advanced search facilities, and visual navigation aids. Based on my experience with Hyper-G, the distributed information system developed at Graz University of Technology, I have the strong feeling that a combination of the three approaches can significantly reduce the "lost in hyperspace" syndrome, and I propose to explore the usefulness of graphical navigation aids to the extent possible.

As an example, Harmony (the X11 client for Hyper-G) supports a "collection browser" which lets users navigate through the hierarchical collection structure of the Hyper-G information base. In the analogy of real-world navigation, the collection browser can be compared to a set of overview maps, which has been organized in a hierarchical fashion. A "local map" can be generated on demand which gives a graphical overview of the hyperlink structure around the current document (like a street map).

Probably the most important navigation aid is "location feedback": the collection browser always shows the position of the current document with respect to the collection hierarchy, regardless of whether the current document has been accessed using Hyper-G's powerful search facilities, by following a hyperlink, or using the collection browser itself. In the real world analogy, this feature would correspond to a GPS device coupled with the set of maps.

In addition, we have recently implemented an "information landscape", which is a 3-D graphical overview of the collection structure. Users can "fly" over the hyperspace landscape looking for salient features, select interesting documents, etc. Having three dimensions, the landscape can display more objects without becoming crowded and can encode more attributes (e.g., type, age, size) than the collection browser in 2-D. For example, it is easy to look over a sea of thousands of objects and pick out the green ones (objects that have been recently modified, for example).

In the future, we plan to implement 3-D visualizations of the both the local map (going 3-D has the advantage that the nodes of an arbitrary graph can be positioned so that no interconnections intersect) and search results (encoding numerous features of the objects matched, rather than just a one- dimensional list sorted by match score). Usability tests will have to reveal whether these visualizations are really useful for end users or just gimmicks.

New Fuel for New Visualizations-- James Pitkow

Focus is rightly shifting away from the constrained environment of within document navigation towards visual abstractions. Yet, these new representations and metaphors are limited by the extent of the properties associated with each displayed object. Efforts like the Dublin Core [10] which recommend a minimal set of meta information for networked objects, are an important first step. Our latest focus has been on the introduction of new analysis techniques and properties of objects accessible via the Web to be coupled into visualization systems like the Navigation View Builder [8] and the Hyperbolic Browser [7]. Additional efforts have been on the development of a scalable hyperlink infrastructure [6] to facilitate visualization systems in the determination of first order structure of objects. Our goal is to facilitate the development of new metaphors by the construction of visualization systems that are based on useful and pioneering meta information.

The Role of Graphics in Future Information Delivery Systems-- William A. Ruh

Next generation information systems will be very different than those fielded today. The stimulus for change is the need for organizations to improve the exploitation of their corporate information assets as well as to effectively exploit the massive amounts of information available from external sources and integrate these two. This information is inherently different than the information around which today's systems are built. Current information systems are developed around the management of reasonably sized, highly structured, record information. New information systems will be built around the management of massive, un- and semi- structured, multimedia information.

One of the four major issues that need to be addressed is the organization, retrieval and exploitation of this massive information base. Is current technology adequate for providing integrated retrieval tools on the user's desktop and organized access to large volumes of data? Can the system search a massive collection in a reasonable time and identify precisely the items of interest to the user?

Graphical, tailorable display of information will be critical to this next generation. Access devices will include hand held computers where there will be no keyboard and maybe even no visual display, only audio. Mapping between mediums and modes, understanding what is appropriate and when are all critical issues. This will require a multi-disciplinary approach for development of graphical applications and interfaces.


Stephen G. Eick

Stephen G. Eick is the Technical Manager of the Data Visualization Research Group at AT\&T Bell Laboratories. His research focuses on extracting the information latent in large databases using novel interactive visualizations. This involves inventing the techniques, developing the software tools, and building an infrastructure to mine knowledge from corporate databases so that it can be put to competitive and commercial advantage. His research group has developed a suite of visualizations including tools for visualizing geographic and abstract networks, software source code, text corpora, log files, program slices, and relational databases, among others. Eick is an active researcher, holds several software patents. He is particularly interested in visualizing databases associated with large software projects, networks, and building high- interaction user interfaces. His educational background includes a B.A from Kalamazoo College (1980), M.A. from the University of Wisconsin at Madison (1981), and Ph.D. in Statistics from the University of Minnesota (1985).

Nahum D. Gershon

Nahum D. Gershon is a Principal Scientist at The MITRE Corp. His work is concerned with data and information visualization, network browsers, image processing, data organization, and analysis of medical, environmental, and other multidimensional data. He pursues research in the use of understanding of the perceptual system in improving the visualization process. He has received his Ph.D. in physical chemistry from the Weizmann Institute of Science and has held position at the Massachusetts Institute of Technology, Johns Hopkins Univ., and the National Institutes of Health. He serves as a member of the Advisory Panel of the Earth Observing System Data and Information System (EOSDIS) and the Focus Group on Visualization and Presentation of the White House's Globe Project.

Joseph Hardin

Joseph Hardin has been the Associate Director for Software Development at the National Center for Supercomputing Applications (NCSA) at the University of Illinois from 1992 to the present. Previously, he was the Manager for the Software Development Group, the coordinator of the Academic Affiliates Program at NCSA, and a visiting research associate at NCSA. His experience includes teaching in the department of Speech Communication at the University of Georgia at Athens and graduate work in Linguistics, Sociology and Communications at the University of Illinois. He was Co-Chair of the Second International World Wide Web Conference: Mosaic and the Web. He was a founder and is currently a Co-Chair of the International World Wide Web Conferences Committee.

Joseph has worked as a consultant in business computer communications, database management, and office organization. He was editor for The Champaign-Urbana Weekly news magazine from 1979 to 1982, where he also wrote the wine column. He has received a number of development grants and awards in the area of scientific software development. He has spoken extensively on workstation tools for computational science and the human dimensions of collaboration technologies in cyberspace. Recent Papers, Presentations, Conference Participation

Frank Kappe

Frank Kappe is the leader of the Hypermedia Group at the Institute for Information Processing and Computer Based New Media (IICM) of the Graz University of Technology, and Director of Software Development at the Institute for Hypermedia Systems of Joanneum Research (a non-profit research organization). As such, he has been responsible for the design and implementation of Hyper-G. He holds an M.Sc. and Ph.D. degrees in computer science, both from Graz University of Technology.

James Pitkow

James Pitkow is a Graphics, Visualization, & Usability graduate student at Georgia Tech's College of Computing. He graduated Cum Laude in Computer Science Applications in Psychology from the University of Colorado at Boulder in 1993. It was during an undergraduate research assistanceship at the Laboratory for Atmospheric and Space Physics, that James was exposed to distributed networked environments, like the Internet and the World-Wide Web (WWW), as part of NASA's Earth Observing System Distributed Information System (EOSDIS). That work followed James to Georgia Tech where he is actively contributing to WWW developments. Besides being the originator and principal researcher for the GVU's WWW User Surveys, James has made contributions in the areas of publishing environments, hyperlink databases, access log analysis, and navigation strategies. His current research interests include user modeling, adaptive interface strategies, and information foraging theory.

William A. Ruh

William Ruh is the Executive Director of Information Technology and Services Division and the Director of the Open Systems Center at the MITRE Corporation. The Open Systems Center has a charter for maintaining and developing MITRE expertise in advanced distributed information systems architectures, commercial product evaluations, corporate standards coordination and research in image and text processing, user interfaces, visualization, collaborature, and multimedia. He has responsibility for providing analysis, design, and consulting expertise in all of these areas to MITRE programs. In addition, William Ruh coordinates the corporation's projects in Open Source information processing and retrieval. He has an M.S. and B.S. in Computer Science from California State University at Fullerton and has held positions at IBM and Franklin Pierce College.


1. Andrews, K. and F. Kappe, Soaring through Hyper Space: A Snapshot of Hyper-G and its Harmony Client, in Proc. of Eurographics Symposium of multimedia/Hypermedia in Open Distributed Environments, Graz, Austria, W. Hezner and F. Kappe, (editors), pages 181-191, Springer Verlag, June 1994.

2. Richard A. Becker and Stephen G. Eick and Allan R. Wilks, "Visualizing Network Data", IEEE Transactions on Visualization and Graphics, Vol 1, No. 1, March, 1995, pp. 16-28.

3. Stephen G. Eick and Graham J. Wills, "Navigating Large Networks with Hierarchies", Visualization '93 Conference Proceedings October, 1993, pp. 204-210.

5. Gershon, N.D., and S.G. Eick, "Visualization's New Tack: Making Sense of Information," IEEE Spectrum, Nov. 1995, pp. 38-56.

6. Jones, R., Pitkow, J., Atlas: Supporting the Web with a Distributed Hyperlink Database, 1995. (submitted WWW4 manuscript).

7. Lamping, J., Rao, R., and Pirolli, P., A Focus+Context technique Based on Hyperbolic Geometry for Visualizing Large Hierarchies, Proceedings of the ACM SIGCHI Conference on Human Factors in Computing Systems, Denver 1995, ACM.

8. Mukherjea, S., and J. Foley, Navigational View Builder: A Tool for Building Navigational Views of Information Spaces, CHI '94 Conference Companion, p. 289-290.

9. Schatz, B.R, and J.B. Hardin, NCSA Mosaic and the World-Wide Web: Global Hypermedia Protocols for the Internet, Science, 265, 895-901 (1994).

10. Stuart Weibel, Jean Godby, Eric Miller, & Ron Daniel, The OCLC/NCSA Metadata Workshop Report, 1995.