Robert Weller has a dream. The Woods Hole Oceanographic Institution oceanographer dreams that one day he will be able to turn on his computer in the morning and transform his office into a global oceanography institute. In his dream, he can bring up data on water temperature, air temperature, wind speeds, currents, air pressure, humidity, salinity, and solar radiation from ocean monitoring devices all across the world--all there at his fingertips and all as fresh as today's catch. Then, without leaving the keyboard, he can analyze this data, borrowing time on a supercomputer that may be halfway across the country. If he wishes, he can link up with colleagues from other institutions to look for patterns in the data and compare the details with predictions from ocean models. Results in hand, he can then instruct the monitoring devices to modify their observations to answer new questions or improve later data.
A dream, yes. Today, scientists like Weller can
use computer networks to pass E-mail or retrieve archived data from
databases, but the sort of instantaneous give and take he envisions
is just not possible. That's changing, however. The tools that Weller
dreams of and many more are on the horizon for scientists in all fields,
thanks to the on going explosion in computing power, ever growing
communications capabilities, and entirely new ways of thinking about
what computers can do. It's impossible to predict exactly what shape
the future will take, but if the computer gurus are correct, at least
one forecast can be made with confidence: Over the next decade or
two, "electronic communities" like the one in Weller's dream
will spring up, linking scientists much more elusely than ever before
with information, instruments, and far flung coIleagues. "There
is genuinely a revolution here," says Bruce Schatz, an information
sci entist at the University of Illinois. "Life is going to be
completely different."
The foundation for that transformation has already been put in place. The Internet, the worldwide system of interconnected computer networks, has created an environment ripe for revolution. There are now 1.7 million host computers hooked up to Internet worldwide (a million in the United States alone), with somewhere between 5 and 15 mtllion individual users, says Vinton Ceff, vice president of the Corporation for National Research Initiatives in Reston, Virginia, and president of the Internet Society. And the numbers are doubling each year, he says.
Although there are no statistics on what percentage of scientists are Internet users, anecdotal evidence suggests that some communities, such as computer science and electrical engineering, have almost 100% participation. Even the humanities, traditionally leery of high-tech gadgetry, are joining in. "There's a sort of virtual university being created [with electronic mail]," says James O'Donnell, professor of classical studies and coordinator of the Center for Computer Analysis of Classical Texts at the University of Pennsylvania. "There's one colleague [at another university] whom I used to speak with maybe a couple of times a year. Now we're in e-mail contact twice a day."
Indeed, electronic mail is often the major reason a researcher signs onto Internet, Ceff says. But once there, he or she finds plenty of other uses. In physics, for example, Paul Ginsparg at Los Alamos National Laboratory has set up several bulletin boards that accept preprints and send out the abstracts to thousands of subscribers, who can download full texts of papers that seem interesting (Science, 26 February, p. 1246). In biology, researchers routinely dial into hundreds of databases containing genetic maps, protein structures, and so on (Science, 11 October 1991, p.201). All told, there are now approximately 50,000 databases available over Internet, Ceff says, with many more coming on line each month.
But this embarrassment of riches has created a problem: How can you find anything in that mass of data ? The database file names are usually nondescriptive, so unless you know what you're looking for, you probably won't find it. Even worse is the difficulty of learning which databases are out there and what information they contain. It's as if you have wandered into a library full of books, but the books are identified only by number, and when you open one to its table of contents, the chapter headings are equally unhelpful.
What's needed is better data about the databases, and that's where the first stage of the revolution is occurring. Over the past 18 months, the Internet library has begun to supply its books with "labels" and "tables of contents" and even to offer indexes of its holdings. Driving this transformation are independent researchers who have developed software tools to help navigate the Internet and then distributed them for anyone to use, says George Brett, director of the Clearinghouse for Networked Information Discovery and Retrieval in North Carolina. Now Internet devotees say they don't know how they ever lived without such aids as Gophers, Wide-Area Information Servers, World Wide Web, Archie, Veronica, and Jughead.
Archie, developed by graduate students at McGill University in Montreal, was one of the earliest of these tools. In essence, Brett explains, Archie is a guide to databases--a regularly updated index of the files available at various computers attached to the Internet that can be downloaded by any Internet user. "Ask Archie, 'Where is Kermit software?' and it will tell you all of the thousands of places to find it."
Archie is aimed at people who know exactly what they're looking for and simply need help tracking it down. But the most popular Internet aid, Gopher, is what Brett calls a "browsing tool." Gopher, created 2 years ago at the University of Minnesota, provides an easy-to-use gateway into "infobases"--generalized databases that can contain not only data but any other type of information, including text and multimedia dis-plays. Installed on an individual infobase, the Gopher software supplies a nested table of contents, so that users can zero in on desired information by making a series of choices. Someone interested in New Zealand, for instance, can connect with a Gopher supplied by the Wellington, New Zealand, city council, pulling out descriptions of culture, geography, or whatever tickled his fancy. To date, some 1100 infobases inhabit "Gopherspace"--a virtual library, available publicly through the Internet, consisting of infobases that can be searched via a Gopher--and the number is growing rapidly.
The ranks of the Internet aids also include Jughead and Veronica. Jughead allows a user to search directly for files in an infobase that is indexed by a Gopher, while Veronica can look for files across all Gophers--it is an "Archie for Gopherspace," l?rett says. The most ambitious of the Internet tools to date is WorldWide Web, which Brett describes as a "hypermedia browser" for a collection of databases. Modeled after hypertext programs, it allows a mouse-equipped user looking through one file to move immediately to a related file in a completely different database by clicking on a highlighted word or phrase. In this way a user can trace a chain of ideas or information through a series of files that appear in computers around the world without worrying about where each piece is located. WorldWide Web has recently moved out of its demonstration stage, Brett says, and is being put to "real work" with some 64 databases included to date.
This trend toward making more and more information easier and easier to get will certainly continue, but a still bigger change is looming, the experts say. Increasingly, the Internet will be used to create "electronic communities"--collections of researchers in a single field who are linked electronically and who share information, instruments, software, and even computing capability.
Virtual laboratories. Some of the first such communities may be the "national collaboratories" touted in a recent report by the National Research Council (NRC).* A collaboratory would, as its name implies, integrate people and resources in such a way that a researcher in any location could hook into the system and do his work as if everything he needed--data, computing power, software, instruments, even other researchers--were right in the same building. "You have on your desk all the tools you need for a class of problems and they're all integrated together," explains William WuIf, professor of computer science at the University of Virginia and originator of the "collaboratory" term.
This sounds very much like the dream of Woods Hole's Weller, as well it should, since the NRC report pointed specifically to oceanography along with space physics and molecular biology as fields that could greatly benefit from collaboratories. The three diverse subjects share a feature that makes them good candidates for electronic collaboration: Each has its vital information spread out over many institutions.
Space physicists, for example, collect a wide variety of data in trying to understand how the sun's radiation interacts with the atmospheres of Earth and other planets. "Historically, these measurements have been sent to different institutions and then the people involved must come together and bring their data to compare," says Christopher Russell, a space physicist at the University of California, Los Angeles, who served on the NRC panel. A space physics collaboratory would give researchers access to all of this data electronically, along with the tools --computers and software--necessary to analyze it. It would also allow scientists at different institutions to hook up over the system in order to examine and manipulate the data together.
Eventuallv, predicts Schatz, another member of the NRC panel, such electronic communities will revolutionize "how people interact with information." To Schatz, the most important part of these communities will not be the fact that researchers can control instruments remotely or access data from dozens of sources at once, but rather that the communities will create a new way for scientists to record and share information and insights. In a prototype community Schatz developed for molecular biologists who study the nematode worm C. elegans (see box), scientists can not only add their own data to the system's databases but also create "links" between different pieces of information. Thus a researcher who notices, say, a similarity between two genes can leave a record of that observation; anyone who later brings up information on either gene will be informed of the similarity. This linkage ability, Schatz predicts, will turn computer networks into something much more than the high-powered library and communications systems they are today.
Since the software is developing so rapidly, the only factor limiting how quickly these electronic communities develop is likely to be scientists' willingness to take advantage of their features, Schatz says. When the telephone was invented, he says, "people wondered why they needed it when they already had the telegraph." With that kind of cultural inertia, electronic communities may not come into their own until people move beyond exchanging messages and scanning databases things that could be done (albeit more slowly) by telephone, fax, and overnight mail and start using the networks to collaborate in new ways. "The people who are carrying out this revolution," Schatz says, are the graduate students," those budding researchers who have grown up with the personal computer and who aren't wedded to old-fashioned ways of doing science. They'll be the citizens of tomorrow's electronic communities, inhabiting a world that to Robert Weller seems much like a dream.
