Tapping into the measureless storehouse of information known as the World Wide Web -- the Internet's most popular navigation tool -- requires access to a computer that is linked to the network. For some, this means having the financial resources to own an expensive personal computer, a modem, the appropriate Web "browser" software, and an Internet connection provided by a commercial service. For most, Web use means being associated with a business or university with an Internet connection. Consequentl y, navigating or "surfing" the Web is something usually not done at home. Most still commute to the workplace first, then get on the Information Superhighway.
A recent survey on the World Wide Web conducted by the Graphics, Visualization, & Usability Center at the Georgia Institute of Technology in Atlanta, shows that most Web use comes from school and t he workplace. The survey found that 51 percent of Web users report that their access comes from the educational sector and 30 percent report access from the commercial sector. Only 28 percent of Web users pay for their access personally.
Usage statistics from the Boulder Community Network (BCN), a recent addition to the World Wide Web, confirm that educational and commercial users comprise a large chunk of total BCN usage, and that nearly all use of the network occurs during typical week day working hours.
Dennis DuBe, a publisher for Boulder's Apple Media Lab, believes that most Boulder Internet users are students with access through CU, or people using the Internet from work. "So that means that the Internet in Boulder is a good market for people who are at work or on campus and not a good market for people who are at home," he said.
To find out how BCN is being used, the network employs a statistical software program called Getstats. As users access BCN and navigate through the information, a log is automatically generated that lists the Internet addresses of users, the time of acc ess, and the files or "pages" they view. Getstats, developed by Kevin Hughes, analyzes the log, produces summaries showing usage over time and ranks the most popular information. Getstats can calculate activity for each BCN page, and activity during a particular hour, day or month, to give an indication of usage patterns. Getstats generated statistics on activity for 10 weeks, from Sept. 12 to Nov. 20, 1994.
Probably the most telling statistic produced was that BCN activity tends to be centered around working hours on weekdays. Monday through Friday activity is slow during the early morning hours and begins to rise around 6 a.m. Between the hours of 6 a.m. and 7 a.m., activity jumps 115 percent and continues to rise steadily, peaking at around 10 or 11 a.m. There is a 25 percent drop in activity at noon and then another steady climb until 3 and 4 p.m., when activity begins to decline into the evening hour s. In other words, activity rises during working hours and drops off during non-working hours, including the lunch hour. "This behavior suggests they're not [accessing BCN] from home. Especially the evening numbers," Dube said.
He also suggested that the sudden spike in activity at 6 a.m. and the tapering off during the hours of 6 and 7 p.m. are evidence of East and West Coast users accessing BCN during their respective business hours. This cannot, however, be verified.
Not surprisingly, BCN activity on the weekends is about half that of the average weekday. On Saturdays and Sundays, activity is low and fits no apparent pattern, with the exception of a big jump at 9 a.m. and 10 a.m.
Internet users are divided into different "domains" according to the kind of connection they have. These domains are indicated by the last three letters of the address assigned to a computer connected to the network. According to "The Whole Internet Us er's Guide and Catalog," addresses of computers used by educational institutions end in ".edu." Anyone accessing the BCN from an .edu address is likely to be a student, instructor or university staff member. Addresses of computers used by commercial org anizations, such as AT&T, end in ".com." Government computer addresses end in ".gov," military institutions in ".mil," network resources entities in ".net," and any other organization in ".org." There are many others, but these are the most common addre ss endings.
According to activity compiled by Getstats, BCN usage mirrors that reported by the Web users survey. More than 28 percent of the accesses came from the educational (.edu) sector and more than 16 percent came from the commercial (.com) sector for a total of 44 percent from the two domains. While this number is smaller than that found by the Web survey, it still represents the two largest domains using BCN. Government domain users make up 5.5 percent of activity, the ".net" domain 5.7 percent, and the other domains make up even less.
There are some user behavior differences across the domains. Activity from educational users, probably including many from CU, shows the same overall weekday pattern of peaking in mid-morning and mid-afternoon, dipping during the noon hour, and taperin g off rapidly after 5 p.m. But Saturday ".edu" activity is just as high as that domain's average weekday activity, with most of it concentrated in the morning.
In fact, the spike in activity on Saturday mornings, mentioned earlier, comes almost entirely from that domain. Dube has an explanation for this anomaly. "Just from looking at this," he said, referring to the 10 a.m. jump in activity, "when would you s ay students get out of bed?"
The heavy Saturday activity by educational users may be due to the fact that most students live on or near a campus where networked computers are accessible even on weekends. Activity from this domain on Sunday, however, drops off by more than 50 percen t from weekdays.
In contrast to users from the educational sector, activity by users in the commercial domain drops nearly 70 percent on weekends. Weekday activity by commercial users follows the general pattern but with a slight rise, instead of a drop, at noon and a m ajor spike at 3 p.m. Weekend activity is very low and sporadic with no discernible pattern. Weekday activity by the government domain is similar to the others, with a very pronounced drop from 5 to 6 p.m. Weekend activity from this domain is almost non -existent. Activity from the military and other organizations (.org) domains appears to be similar but so minimal that larger patterns are hard to determine. It is clear that users from these domains leave their computers when they leave work for the we ekend.
BCN activity coming from one kiosk on the bridge of the Boulder Public Library is about 3.5 percent of total BCN usage. While that number seems small, it is larger than the amount of activity coming from the military and other organization domains combi ned. If that is an indicator of public kiosk use as a whole, than the BCN has been relatively successful in supplying access to the public.
Activity from the library kiosks, naturally, is a function of library hours, so there is none after 9 p.m. on weekdays, or after 6 p.m. on weekends. The bridge kiosk gets more use on Saturday than any other day, and activity on weekdays and weekends ten ds to peak at about 1 p.m. or 2 p.m. It seems appropriate that this activity, on a terminal located near an espresso bar and a gift shop, occurs nearer the lunch hour. There is no way, other than old-fashioned observation, to know for sure.
As the Internet has gained popularity, companies have emerged to supply individual home users and businesses with access. Colorado SuperNet is one such company supplying access to users across the state. Colo rado SuperNet allows users to choose their domain names or gives them the default of ".csn.net." BCN analyzed activity from that domain name for some interesting results. Colorado SuperNet activity, with both home and business users, appears more orient ed to home use than general BCN activity. Activity rises from about 5 a.m. to about 10 a.m. then drops off. It is heavy in the mid-afternoon but continues to be so into the evening, when it peaks at 8 p.m. before dropping off into the late night hours.
Prodigy, a commercial on-line service with home and business users, reports similar user activity. Brian Eck, director of communications for the service, said activity is concentrated in the early mornings and early evenings. Spokespeople at Prodigy do n't know the ratio of home to work subscribers, but they say 35 percent either own their own business or operate one from home.
Dube said that so many work and school-related Web users means that there is still a good home market for traditional print products. "This is particularly good news for newspapers," he said. In fact, the tendency for activity by home users to be cente red in the mornings and evenings is similar to the way most daily newspapers are read. Studies have found that the most popular time for newspaper reading was after the evening meal.
If they were cheaper and more accessible, on-line services might one day take the place of the newspaper in the home. The newspaper industry has, in fact, experienced decreasing readership in recent years, partially due to the competition of other media , including on-line services. Many newspapers, such as the Gazette Telegraph in Colorado Springs, have begun producing on-line versions as a supplement to their print products in order to try to win back some readership. Phil Witherow is director of res earch for the Gazette Telegraph, which produces an electronic version of the paper called GTOnline. "We think there are a number of people in the high-tech area who have dropped out of newspapers and we thi nk [GTOnline] will bring them back," Witherow said.
According to statistics on use of GTOnline, activity is similar to that of BCN. "It seems to be during the day," Witherow said. "Most of it seems to be during the business day." This holds true even though, according to a survey conducted by the paper, half of El Paso County and nearby Teller County adults own a personal computer, a significantly higher penetration than the nation at large.
But many, particularly older Americans who grew up with print media, are reluctant to get their news from a computer screen. The failed videotex experiments of the early 1980s, such as Knight-Ridder's Viewtron, are testimony to this fact. According to Roger Fidler of Knight-Ridder's Boulder-based Information Design Laboratory, a screen, for most, is not as pleasurable to read as ink on paper. This, Fidler said, is because co mputer screen technology is based on that of television where there is a barely perceptible but constant flicker, and the resolution, at 70 dots per inch (DPI), makes reading difficult. "Most people will tell you they don't want to read a video screen," F idler said.
His lab is exploring technology for an electronic tablet or "flat-panel," much like Apple's Newton, that would combine the convenience and readability of a print newspaper with all the functions of a computer. The tablet screen will have higher resoluti on and a constant image that doesn't require refreshing, Fidler said. But early prototypes of the technology won't be available until the end of the decade. Fidler said this tablet will be simpler to use than the Web. "We have to do things in a way tha t is quickly understandable and doesn't require a manual," he said.
Until Fidler's vision is reality, many think computers are better-suited to doing searches and reading short but frequently updated news briefs, while ink on paper is still easier and more pleasurable for reading longer blocks of text. One recent study c onducted by the research team of Clark Martire & Bartolomeo Inc. found that most people will not spend 25 minutes reading news and features from a computer screen the way they will from a newspaper.
In fact, the most frequently accessed information on the Boulder Community Network supports these findings. To see what was being accessed, the top ten "destinations" on BCN were analyzed for one week in November, 1994. Nearly all documents on the BCN a nd Web at large have Hypertext links to other documents. The surfing metaphor is an appropriate way to describe how the Web is used. Users go in and, with some direction and a lot of serendipity, move about through the information looking for something interesting or useful. Many of these pages merely serve as indexes to other pages and are not places to stop for information. A "destination," then, was defined as a page with some type of useful information that had no links to further documents other than a return to a previous page or index.
The most popular destination on BCN is called the "Public Information Format," and contains a list of employment and career-related services, such as the YWCA, the Boulder Department of So cial Services' Job Opportunity and Basic Skills Program (JOBS), Project Self-Sufficiency and others. The list includes short descriptions of who qualifies, what services are offered, what they cost and how to reach them. In other words, BCN's most popul ar destination is a list that prompts readers to use the telephone to find more information.
Five of the top ten -- the second, third, fifth, sixth and eighth destinations -- are lists of local happenings provided by the Colorado Daily newspaper. They are, in this order, a movie s chedule, weather, concerts, arts and entertainment. Each of these consists of short, one paragraph briefs with times, places, costs and the like. These are not Colorado Daily articles, or reviews, but simply listings.
The fourth most popular destination relates to another advantage that on-line information services have over print: interactivity and communication. This destination allows users with the appropriate software to send electronic mail to BCN staff, as well as local and national government representatives, including the president. To use this feature, users must have the most recent version of Netscape, a Web browser that has a "forms" function. T his feature allows users to access the page, type in a quick message along with their own e-mail address, choose a receiver's address from a list, and send the message. The same software allows users to post messages to a discussion group on BCN. Howeve r, this group is not yet very popular, probably because it is still in the BCN test area and hard to find. The on-line Campus Press, a student-produced newspaper on BCN, has also recently added a discussion area that is likely to be popular among students.
At the seventh destination, users are finally getting some actual news. This page is called "City News Briefs" and contains just that: short paragraphs of information about local happenings such as "Meeting Management and Facilitation Training for Neighborhood Leaders (Leaders will learn how to create meeting environments that encourage participation and strengthen dialogue among participants.")
The ninth destination is a list, produced by the Boulder County Business Report, of recent hirings and promotions in local businesses.
Finally, the tenth most popular BCN destination, called "Accessing BCN," describes the type of connections and software needed to access the site, as well as a list of available public kiosks. This highlights the fact that BCN and the World Wide Web are relatively new on the scene, and people, including BCN providers, are still learning what the system is all about.
In short, BCN is being used mostly to get quick information and not to read longer articles, though they do exist in places like the student-produced Campus Press. Added together, the Cam pus Press pages make up the most frequently accessed of any other group of information on BCN. But users seem only to be going in, surfing around, and getting out without stopping to read the articles. In fairness to the Campus Press, many of its pages, which rank high in usage, contain short article summaries along with links to other documents and, therefore, were not considered destinations.
Even accesses to the first of the most popular BCN destinations made up less than one percent of total BCN usage.
Though newspapers are said to be better-suited to long articles, one study showed most people don't read most of them anyway. The Poynter Institute for Media Studies conducted a 1991 study of newspaper reading habits by using a special camera mounted on the heads of readers that could track their eye movements as they read a newspaper. The results, published in a book called "Eyes on the News," found that only 25 percent of the text in newspapers was being looked at or "processed," compared with 80 per cent of the artwork, 75 percent of the photos, 56 percent of the headlines, 52 percent of the advertising, 31 percent of the news briefs and 29 percent of the photo cutlines.
BCN also has photos and "artwork" in its on-line galleries, but they are not high in the ranking probably because photos take a long time to come up on the screen and can be frustrating to wait for.
The Poynter Institute findings show that most readers might be ready for reading from a screen after all. The Information Design Lab's Roger Fidler notes that the general lack of newspaper readership shows that the information rich and information poor have been around for years. "There's always been a gap there. Let's face it, a lot of people don't read newspapers," he said.
Until Fidler's version of an electronic newspaper hits the market, BCN will continue to try to provide access to this growing world of information for anyone who cares to take a look. Most, however, will have to leave home to do so.