Early explorers and navigators of the uncharted areas of our planet had to progress and advance their expeditions with little assistance, but followers were more efficient, had less risk, and were also more effective, as their maps were more informative of alternative paths, dangers, and conditions. When surfing the web, each expedition is almost as novel as any previous one. The mechanisms to provide the ability to step back, to retrace our steps, and to prevent us from going to where we do not want to go, are primitive.
We need to provide tools for the user to find alternative paths, or shortest paths, besides adding interesting URLs to our unmanageable list of favorites, so all pages are one-click away. In our real world, travelers, pilots, and sailors need maps to guide their journey. Maps provide a capacity to plan routes, to retrace steps, to evaluate alternatives. Our hypothesis is that maps can be useful in guiding web journeys.
In fact, recent web-masters are adding site-maps to their servers. Examples are: Microsoft WebMapper, IBM's Mapuccino, and Lotus' WebCutter. These are also linear listings or perhaps simple diagrams, with hierarchically structure of the many documents in a site, but are still very limited in visually conveying information and a sense of space and proximity. They are even less successful in displaying which documents and pages refer to others and which are the existing alternative paths. Finally, they limit their information to one individual site; they provide no map of the world beyond.
1.Current Web Browsers & Site Mapping Techniques
Hypertext is a type of database that has active cross-references and allows the reader to "jump" to other parts of the database as desired. This makes the reading (and writing) process non-sequential. A hypertext database can be conceptualized as a network of nodes and links, where documents are the nodes and the links are cross-references. We name these documents as hyper-documents, and these cross-references as hyperlinks. The underlying data model of this hypertext network can be viewed as a very complex graph; it is a partially unknown huge graph.
The World Wide Web (WWW) is a kind of hypertext database. The amount of information now available through the WWW has grown explosively. An increasing number of tools are also available to assist the user to manage and access information on the WWW. One of the key requirements for a WWW navigator is to maintain the user's sense of orientation and facilitate navigation within the context of the total information space.
The modern web browsers, such as Netscape and Microsoft Internet Explorer, can effectively assist the user to manage and access information on the World Wide Web. They excited people's interest because of the structure of hypertext that provides users with an effective and convenient way to move in cyberspace. This is done by clicking on a series of hyperlinks embedded in hyper-documents.
However, these web browsers can not give users a visual "map" to guide their web journey. They do not provide a sense of "space" while the user is exploring the (cyber) space; instead they only give a series of lists (bookmarks and history lists are at most a one-dimensional graphical presentation). This is mainly because of the difficulty of constructing such a huge, complex, and dynamic map with a (virtually) unlimited number of nodes and edges.
Most existing visualization techniques emphasize "site mapping". That is, they try to find an effective way of constructing a structured geometrical map for one web site (a local map). For example, a hyperbolic tree Overview Diagram has been developed by Inxight Software Co. However, the "site mapping" technique only maps a single web site, and can only guide the user through a very limited region of cyberspace. It does not help users in their overall journey through the cyberspace.
2.Overview of Visual Web Browser
The Visual Web Browser is expected to be the next generation tool for Web navigation. It views the whole cyberspace as one graph - a huge and dynamic growing graph which is partially unknown. At any time, a tiny subset of the huge graph is known and a picture of this subset is displayed on the screen. Exploration of the huge graph proceeds by changing the subset of the huge graph. This allows the user to visually browse through the Web with a sense of "space". The Visual Web Browser explores the huge graph by quickly tracking the subset of the huge graph based on the focus of the user, and provides the user with a dynamic visual map for guiding the web journey.
The Visual Web Browser is like an on-line geographical map. Geographical maps do not offer you information on every geographical feature. Some of it is filtered, thematically, or by the scale. Too much in a map makes it unreadable. While traveling with the aid of a map, typically we focus in the region or neighborhood that currently surrounds us. That allows us to decide if a turn at the next intersection is needed. So our Visual Web Browser maps the neighborhood of the web that we are currently in, filtering some of the information.
The Visual Web Browser displays the links and paths from document to document. This is a visualization that conveys the sense of space and the sense of location with spatial dimensions. A sense of proximity is represented by placing nodes in proportion to the measurements of time/cost of retrieving a page. The visualization is animated to preserve the sense of orientation. As we move in cyberspace, new nodes are smoothly added into the map and others far in diametrical direction fade away.
3.Visual Web Browser architecture
The Visual Web Browser is made up of several components. These components can be described in Figure 1; details are below:
Figure 1: The architecture of Visual Web Browser.
An on-line animated visualizer: It automatically displays a sequence of web sub-graphs with smooth animation following the user's orientation. It adapts our on-line exploratory visualization technique. It does not pre-define the geometrical structure of a specific web site (a part of cyberspace); instead it incrementally calculates and maintains a small visualization of a subset of cyberspace on-line. This feature enables the user to logically explore the entire cyberspace without requiring the whole structure of the cyberspace to be known.
A fast accessible linkage server: To quickly obtain the linkage information (the neighborhood of the focused URL) for dynamically creating the visualization, we need a fast accessible linkage server. This server is separated from the main web server. It can quickly find and produce the entire neighborhood of a given focus node. In its basic operation, the server accepts a query of an URL of the user clicked focus node, and then returns a list of all pages that are pointed to, from the page representing the focus node.
An information filter: To reduce the complexity of the web graph, we provide a filtering mechanism to remove unnecessary information (links and nodes) from the visualization, and only retain the essential part of the web graph. This simplified visualization is a tree structure. We then use the on-line animated visualizer to maintain the user's orientation for web exploration.
4. Online animated visualizer
The visualization technique adopted in the on-line animated visualizer, known as OFDAV provides a major departure from traditional visualization methods. It allows the user to visualize the entire cyberspace that is available through the hyper-media system, but does not require the whole web graph to be known. It assumes that the amount of data that can be effectively displayed at one time is only a tiny subset of the available cyberspace. So it does not predefine the geometry of the whole visualization at once (the user can navigate logically); instead it incrementally calculates and maintains a small local visualization on-line, corresponding to the change of the focus page of the user.
We use a force-directed algorithm to draw this sub-graph and the logical neighborhood tree around this sub-graph. The logical neighborhood tree of the focus nodes gives the user a clear idea of where they are and helps them to decide where they should go next. It provides a direct visual aid guiding the user for further navigation.
We use multiple animation to guide the user between views; they make the transitions naturally and smoothly. The change of the user's mental map can be naturally followed with the animated motions of the updates. In the user's visual sense, there is only one animated image. This greatly reduces the cognitive effort in re-forming the user's mental map after each transformation. We also adopt a graphical history tail that contains a sequence of previous focus nodes. It traces the sub-graphs that the user has visited and assists in backtracking through the web graph. In addition, the transition from the picture of one logical frame to the next should preserve the mental map, that is, the difference between successive drawings should be small enough that the user perceives the transition to be smooth.
5. Remarks
This article introduces a new approach to web browser design. This design integrates three components: the on-line animated visualizer, the linkage server, and the information filter. The on-line animated visualizer provides the user a sequence of animated drawings of web sub-graph that dynamically follows the user's focuses in the process of navigation. Each web sub-graph is formed by using the linkage server to search the focus node's neighborhood and using the information filter to remove some unimportant nodes and edges for navigation.
The Visual Web Browser provides users a dynamic visual map to guide their web journey. This map gives the user a sense of "space" while the user is browsing and navigating the World Wide Web. The layout of web-graphs generated automatically looks nicely symmetric and well spread out.
The practice of Visual Web Browser has shown that it is an effective tool for web navigation and information discovery of any type of large data spaces. The experience of using Visual Web Browser has also demonstrated that it is worthwhile, and is a good technique that can be applied in any sub-area of software engineering and data engineering for the general purpose of information discovery.
Dr. Mao Lin Huang
School of Computing Sciences, University of Technology, Sydney, NSW 2007, Australia. Email: maolin@socs.uts.edu.au
Editor's note: A more robust version of this paper was presented in the Poster Proceedings from the 8th International WWW Conference. The author was one of many conference presenters developing next generation web browsers.
SLA Home Page | Join SLA Now | Feedback |
Copyright © 2000 SLA. All rights reserved.
RSS
Feedback form
