Spinning the World Wide Web

Spinning the World Wide Web

As the Web evolves from an information system to a staple of our culture, experts offer different views on the future of the phenomena in the next decade

By John Makulowich

What Tim Berners-Lee, the creator of the World Wide Web, defined in 1992 as a "distributed heterogeneous collaborative multi-media information system," has evolved into an industry and a social phenomena.

One measure of growth can be seen in the number of Internet service or access providers operating in the United States. Boardwatch magazine, Littleton, Colo., counts 3,100, twice the number from last year. According to the monthly magazine, the average Internet service provider has 2,200 customers.

Another measure of the Web's popularity and penetration is Web advertising revenue. It reached $24.8 million in November, up nearly 30 percent from $19.1 million in October, according to estimates from Electronic Advertising & Marketplace Report, a bi-weekly newsletter published by Cowles/Simba Information. For 18 leading Web sites, the revenue from ads totaled $16.1 million in November, up from $14.3 million in October.

Yet a third measure, the most obvious embodiments of the technology, as well as the competitive challenge that it represents, can be seen in the Web browsing clients and supporting software from Microsoft and Netscape. Not only does Microsoft have its Internet Explorer 3.01 with a myriad of productivity tools, but it recently introduced its revamped online service, the Microsoft Network, with an impressive new front end featuring enhanced graphics, audio and video.

Hardly to be outflanked, Netscape just released its newest offering, Netscape Communicator, an e-mail-groupware-browser suite of five to eight tools (depending on the version) for what is no longer just browsing, but working with the Web. Both Netscape's 22mb package and the 42mb MSN product from Microsoft raise anew the threshold of expectations for end users and industry observers.

The question on the minds of most developers and systems integrators, if not end users, is where is the technology heading? While no reasonable person would hazard a prediction, there are two viewpoints worth considering.

Enter Kenneth C. Green, a visiting scholar at Claremont Graduate School, Claremont, Calif., who just completed his 1996 National Survey of Information Technology in Higher Education, and Donna K. Harman, manager of the Natural Language Processing and Information Retrieval Group at the National Institute of Standards and Technology, who has organized the Text Retrieval Conference or TREC, since its founding in November 1992.

National Higher Education Survey

In his recently completed national study of the use of information technology in higher education, Green found that "instructional integration and user support are the two most important information technology issues confronting American colleges and universities over the next two to three years." The survey compiled replies from computing officials at 660 two- and four-year colleges and universities across the United States. It has been conducted since 1990.

The director of the study also found that two-thirds or 67 percent of all undergraduates have access to e-mail and the Internet, up from 60 percent in 1995. And four-fifths or 79 percent of the responding campuses have an institutional Web presence, up from 55 percent in 1995. In an important trend, over one-sixth or 17.5 percent of the respondents have a formal plan for the role of the Internet and the Web in distance education, up from 12.5 percent last year. And more than half or 56.8 percent have a formal plan for using the Web in off-campus promotion compared to almost two-fifths or 38 percent in 1995.

More important for the companies of tomorrow that will hire the students of today is Green's assessment that colleges confront growing expectations from students across all disciplines that technology will be part of the learning and instructional experience.

On the other side of the spectrum in a national government laboratory run under the guidance of the U.S. Department of Commerce, Harman plies her trade as an engineer interested in issues surrounding text retrieval.

Among the many reasons why this research is important is the fact that search engines are among the top sites on the Web in the amount of traffic they generate from users seeking information across the range of knowledge and data.

They also rank high on ad revenues. Yahoo was first recently with $2.2 million for the month of November, up slightly from $2 million in October. Infoseek was the second most popular site among advertisers, attracting an estimated $1.7 million in November. WebCrawler was third with an estimated $1.5 million in advertising revenue.

Text retrieval is also a critical part of the Web's raison d'être, that is, resource discovery and collaborative work among professionals.

Keeping the current state of search engine technology in perspective, Harman points out that these tools are a very small piece of the Web searching mechanism.

"You could have the most terrific search engine and you would only have one-fourth of the puzzle. You also have to ask what the specific search engine managers pick to use as their indexable data. Each of today's sites have specific things they go after. You have to ask yourself which particular sites do the particular engines index? Once you know that, you then ask how they chose to index it? Do they index only the title, the first paragraph, the full text? The third component is the interface, which is clearly important to getting and keeping the user on the Web and on your specific page. Looked at this way, the search engine, the algorithm, is the least important of the lot," explains Harman.

TREC-6, scheduled for November 1997, will again bring together information retrieval researchers to discuss their system results on large test collections. The groups participating in the first five TRECs represent the Who's Who of the online retrieval industry.