EDUCATION VIA ADVANCED TECHNOLOGIES

Final Report, MIT Committee on EVAT


How is the Web Likely to Evolve?

The committee believes that the Web will, over the next few years, evolve in ways that are in part predictable.

Hardware

Hardware for the Web consists of personal computers that run browsers, workstations (or personal computers) that run servers, and communications links.

Personal computers will improve in speed, storage, and display at rates similar to what has happened during the past decade, while not increasing in price. The technology that underlies these computers is microelectronics, which is expected to continue its astonishing rate of progress for at least ten more years.

Personal computers will also improve in portability. In the near future several campuses (including MIT) will be wired for cellular radio connections at data rates that are useful for the Web. We can expect all our students, a few years hence, to have portable, book-size computers with wireless Internet connection. Perhaps the principal limit to portable computers will be battery technology.

Workstations that run Web servers will similarly get faster, have more data storage, and be able to support more complex Web services.

Internet bandwidth will increase dramatically. Currently most Internet connections are at rates below 2 Mb/sec, although there is a high-speed spine within the continental U.S. which achieves 34 Mb/sec. In the future, the amount of traffic will require upgrading many Internet links and as a consequence full video will be possible. The internal MIT networks support 10 Mb/sec traffic all the way to most individual desks.

High-speed Internet connections will come to most citizens at home, primarily through cable TV companies. The Web requires high-speed communications from the server to the browser, but not the other way around. This asymmetry matches perfectly the typical hardware configurations of cable TV.

Software

Currently there are only a few Web browsers. In the future the committee expects there to be hundreds. Operating systems for personal computers will incorporate browsers seamlessly, and special-purpose browsers will be contained in other software. Software programs that requires up-to-date data (e.g., financial software) will have built-in access to data bases on the Web.

Server software will undoubtedly advance, but even more interesting will be the development of a large number of special-purpose software programs, running on server machines, to perform specialized interactive Web services. These programs will be important for educational uses of the Web.

An extension currently under development is Java, which is a language to direct the downloading, from a server, of executable programs to be run on a browser. This concept could change the character of the Web dramatically.

Authoring tools will be available to permit easy Web publication in many fields. These will take advantage of field-specific knowledge. Software will be available to enable faculty members to do little more than specify the content in relatively simple plain English, and then get high-quality Web pages for a variety of uses. We will have available automatic conversion between HTML or its successors and a wide variety of other data formats. It will be as easy to publish on the Web as it is to write a paper or a book chapter today.

A key problem may be lack of industry-wide standards. The formatting capabilities of HTML are astonishingly primitive, and already some Web browsers have implemented nonstandard commands for such features as text centering, font control, and the like. Apparently the developers of these browsers hope to differentiate their products by means of incompatible features.

Media

Currently the Web supports many media: text, graphical images, sound, and video. However, complex images, audio of more than a few seconds duration, and all video require so much data transport that there is a significant delay in downloading. The data may not fit on the browser machine, and its arrival invokes awkward helpers, or "external viewers." Thus the Web is most efficient for pages that consist of text and in-line simple graphics, with other media linked but not automatically called. In other words, content in the other media is there but the user must ask for it, presumably with an understanding of the cost. Also, the technology of incorporating hyperlinks into these media is still unknown or primitive, so that these media can be referenced by pages of text and images, but not the other way around. This means that audio and video are not really first-class Web media. In the future these limitations will probably disappear.

All it takes for the Web to support a new medium is for a special-purpose helper application and a convention for the format of the information. Web browsers are designed to accommodate novel helpers easily. There will probably be a multitude of incompatible formats and helpers, with the effect that Web publishers will never know the exact capabilities of their readers. The possibility of an unstructured jungle exists.

Content

Right now the material that is available on the Web is of variable quality. Some of it is well done, and some is junk. Some is artistically pleasing but much is ugly. Some data is maintained up to date but some is no longer current. Some data is accurate but some is not. Broken links are common. In other words, the Web is pot luck -- users must be cautious in dealing with what is there.

The Web's variable quality of content may be due to the fact that the "early adopters" of this technology were those who were enamored of the technology itself. At MIT the students caught on to the Web long before the university administration. In the future, as more organizations establish a presence on the Web, they will maintain their Web pages with the same care and attention paid to other public relations. Thus the content on the Web will be more reliable, and users of the Web will be able to infer reliability from recognition of the authors or organizations responsible.

Not surprisingly, much of the content on the Web today deals with its associated technologies. For every Web page dealing with botany there are probably a hundred dealing with computers. However, as the Web proves its usefulness, and as it becomes more pervasive in society, its content will expand to mimic society itself. Thus it is reasonable to expect within a few years a vast collection of resources, in all fields of interest. In its breadth the Web will rival public libraries, and its content will be an irresistible educational resource.

Public Acceptance

Today the Web is perceived by the general public, at least those who know of it, as a toy, or a tool only for the elite. Even many computer-literate people are not yet aware of it. Many people had this same attitude about computers before the widespread deployment of personal computers. A similar attitude about e-mail is now disappearing as the general public has, through services such as Compuserve or America Online, experienced its benefits. This view of the Web will change as access becomes easier and more people find Web surfing to be fun and useful.

What We Cannot Predict

The features of the Web that account for its popularity are likely to be incorporated into new systems that will obsolete the Web by doing everything it can, and more. We do not know what additional capabilities these systems will have. However, it is unlikely that such new systems will abandon the successful features of the Web. In this report we are careful to note that our recommendations are not contingent on the Web itself being the system of choice, but apply equally well to new, yet-to-be-developed systems so long as the crucial advantages of the Web are preserved.
This page revised June 1, 1995. Your comments about this report are welcome.
To the Table of Contents. Copyright (c) 1995 Massachusetts Institute of Technology