(FORTUNE Magazine) – IF YOU ARE among the hundreds of thousands of Americans who have recently begun using ''on-line databases,'' you doubtless have a strong sense of doing something futuristic. The explosion of these electronic databases--to which personal computers gain access via telephone lines--has created a sensational new growth industry. Something like 2,400 on-line databases exist already, and scores of new ones are created every month. They have begun to transform the way executives and professionals get vital information; for many of those people they have also come to look like one of the more interesting adult toys of the Eighties. We are still learning about the economic consequences of putting so much information at the instant disposal of millions. In one sense, on-line databases aren't new at all. Terminals supplied by Quotron Systems, Automatic Data Processing, and Bunker Ramo Information Systems have been giving brokers instant stock market information for some 20 years. Mead Data Central's Lexis terminals, which enable you to look up regulations and judicial opinions, have been in virtually all major law libraries for half a dozen years. Mead Data's Nexis terminals, offering on- line access to articles in scores of publications, have been in many other libraries since 1980. (Mead Data is a subsidiary of Mead, the paper company.) What's new about on-line databases is their link to personal computers. Two conjoined events changed everything. The first was the microcomputer explosion: 17 million are now in U.S. homes and offices, up from ten million a year ago and five million the year before. The second event was the proliferation of the Hayes Smartmodem and other ''intelligent'' modems making it relatively cheap and easy to connect personal computers to telephone lines. A growing fraction of PCs, probably around 10% nowadays, have modems enabling them to search around in databases. The customers for the databases have been changing. Several years ago virtually all were research librarians; today a significant number are the professionals or business executives who are actually going to use the information--what the database industry calls ''endusers.'' (The industry has a tendency to employ one word where two would have done just fine.) In general, the industry prefers to sell to endusers, who tend to run up heftier time charges, at anywhere from $5 to $75 an hour, than do librarians. ''Our strategy is to follow the professional home,'' says Mead Data marketing V.P. Paul Nezi. Even Quotron has lately been selling financial quotes to people with micros. Its new Quotdial service, for which a moderately heavy user might pay $150 a month, gives you real-time quotes on your own PC. What's available on-line to an enduser these days is awesome: ) The NewsNet database collection puts more than 200 specialized newsletters on-line, enabling subscribers to update themselves instantly on the latest developments in tobacco exports, small-business tax problems, organized religion, and you name it. ) The Telerate databases, which concentrate on credit market and currency quotes, also offer up-to-the-minute information on ski conditions, the jet aircraft market, and Japanese baseball scores. ) Legi-Slate, a database owned by the Washington Post Co., can be used to check the current status of any bill or pending regulation, or to check the voting record of any member of Congress since 1979. You can create a customized list of critical votes and generate instant rankings of different Congressmen according to your priorities. ) Weatherscan International offers weather data, updated every minute, from any of 10,000 worldwide reporting stations. ) Nexis offers the largest sample of media databases, enabling you, for example, to search out and retrieve the full texts of news stories and articles from the major wire services, ten newspapers (including the New York Times and Washington Post), and 48 magazines (including FORTUNE and all other Time Inc. publications). Other entries include the financial statements of 4,000-odd public corporations, the Congressional Quarterly, the Federal Register, specialized trade publications like Platt's Oilgram News, and such oddball collections as the entire record of the Baldwin-United bankruptcy case. ) Dialog Information Services, a subsidiary of Lockheed, is by some measures the largest database empire of all. It offers subscribers access to over 200 databases and plans to add at least 30 a year in the near future. (It added 43 in one recent year, but Dialog President Roger Summit says that pace proved ''perturbative.'') The present collection contains over 100 million records, with citations to articles in 10,000 different journals, both those figures being unmatched by other distributors. It offers, for example, abstracts of just about every book or article on issues in philosophy published in the U.S. since 1940. It has on-line citations for every book in the Library of Congress. The service, which has about 55,000 subscribers, figures to become much more popular as a result of a software package that came on the market a year ago. The package, called In-Search and developed (with Dialog's blessing) by Menlo Corp. of Santa Clara, California, makes all the databases far easier to use than they had been. Dialog also has one large limiting factor on its growth: about 70% of its databases offer only abstracts and citations, not full texts. It can instantly serve up a staggering amount of information on what has been published on almost any subject, but at the end of the typical Dialog search you will probably want to go off and find the printed text. Dialog management agrees that eventually all databases, including its own, must deliver full texts. A major question about the database business is whether the players who will dominate it in the years ahead are those already on stage. You might suppose that the continuing proliferation of databases means it's an easy business to enter; you might be fortified in this opinion by the fact that little of the data now on-line is exclusive or proprietary. In fact, there are significant barriers to entry. Economies of scale are huge in the business, which is why outfits like Mead Data and Dialog have found it natural to keep adding more and more data to their collections. Mead Data has annual revenues estimated at around $130 million, vs. a bit more than $50 million for Dialog. Quotron, with an estimated $190 million, is first in revenues. Having achieved this scale, all three outfits have tremendous advantages over newcomers. All three are profitable but seem to be pricing at levels that emphasize growth rather than fancy profit margins. If you spend much time talking to people in the industry, you can tune in on several visions of the industry's future. In one, just about all information everywhere eventually gets to be on-line, and just about every company and every educated individual is a customer. This vision raises several questions. One is whether, in an electronic era, it will ultimately make sense to keep huge gobs of data on-line. An interesting alternative is the laser-scanned optical disk, whose data storage capabilities will clearly be enormous; it is already possible to get the equivalent of 12,000 pages or so on one disk. So possibly people will prefer having their own electronic libraries to having the ability to search in on-line databases. Assuming that on-line technology will prove cheaper and prevail, questions remain about the technology by which data get into an electronic database. When a database consists of material that has already been published, and when the publisher routinely prints from photocomposition tapes, everything is a relative breeze: the publisher simply transmits his text to the database distributor via phone lines, and the technicians have only to adjust some formatting codes in the tape. Since most major publications do in fact use tapes to drive their presses nowadays, it's relatively easy to create or add to databases from their text. Unfortunately, a lot of text that's extremely marketable isn't easy to get into electronic databases. Only a negligible fraction of the 20 million books in the Library of Congress have been printed electronically. Our judicial system still gets by without much electronic printing; about 90% of Lexis has to be keyed in by hand, one character at a time. Furthermore, the publications now using photocomposition tapes have done so for only a few years, so that getting back issues into a database is tough. THE FOLKS at Nexis insist that the keyboarding problem is not insuperable and that if only the demand were there they would be delighted to key in older issues of the New York Times, for which full text is available only since June 1980. (Washington Post text is available since January 1977.) Most Nexis-Lexis keyboarding is done offshore, typically in Taiwan or the Philippines, but is still relatively expensive: it costs the company around 70 cents per 1,000 characters. (Somewhat mind-blowingly, much of this work is performed by keyboarders who don't understand English. Nexis executives say they prefer such workers, who are less likely to be distracted by the contents of the text they're replicating.) If the company wanted to provide older issues of the Times or Post, one Nexis official estimated, its keyboarding operation could move back through time at the stately pace of perhaps two years every six months. The process might become cheaper and easier with improvements in scanning technology. For many publications, the database operators already have a choice between keyboarding and electronic scanning. The two have been roughly competitive in price and accuracy, but Kurzweil Computer Products of Cambridge, Massachusetts, a major manufacturer of scanners, says its new Model 4000 will eliminate a lot of offshore keyboarding. Still, scanning technology has a way to go. The scanners work only with sharply defined characters; the operators must often tear the printed page into single columns, which must be perfectly aligned, and even at that the machines make errors.

The technologies associated with retrieving data via random access have raced far ahead of those associated with getting the stuff into the databases. The gap is dramatically evidenced at the Library of Congress. The library has been keyboarding like mad since 1968 and now has its entire card catalogue on-line. It still has only the most tentative notions about putting books and other text on-line. Yet the library has already taken possession of the famous ''jukebox,'' a machine that can hold 100 of those amazing optical disks and randomly access any piece of data on any of them. Not everyone foresees a huge market for raw data on-line. Many people in the industry believe that once you get away from time-urgent data (price quotes, for example), the market will be relatively limited and will be increasingly tied to some kind of marketing--that is, to arrangements linking the on-line data to business and consumer transactions. Some database distributors are plainly interested in this concept. Dow Jones News/Retrieval, which offers access to a range of news services (including the text of recent Wall Street Journal articles) and stock market quotes, will soon offer a discount-broker trading service as well. Both Source Telecomputing (a subsidiary of the Reader's Digest Association) and CompuServe (owned by H&R Block) have a fair number of subscribers who use them principally for communicating with one another via electronic mail and teleconferencing arrangements. Both already have discount brokerage services and are beefing up their capabilities in other kinds of electronic shopping. CompuServe has a new ''Electronic Mall,'' * in which Sears, Waldenbooks, and dozens of other companies are selling to micro users. Dow Jones, the Source, and CompuServe also carry the Official Airline Guides, which tell you who flies where all over the world and also quote North American prices. You can buy an airline ticket with the Source or CompuServe and a credit card. The Source also allows you to make hotel and car-rental reservations. The Securities and Exchange Commission is developing an on-line database that is very much intended to affect transactions. Its EDGAR project (that stands for Electronic Data Gathering, Analysis, and Retrieval) has been experimenting with procedures for putting all required financial disclosures on-line; more than 100 companies have been transmitting reports to the SEC electronically in recent months. Within two years the commission hopes to have all public-record filings accessible by computer. Any investor with even a modest home computer like the Commodore 64 will be able to call up any financial report on his screen. An SEC statement this fall solemnly cited a survey in which 32% of individual investors said they expected EDGAR to improve their financial performance. Those investors will possibly turn out to have been optimistic, but it's reasonable to suppose that EDGAR will make the financial markets more efficient--that is, by more rapidly disseminating information to more and more investors, it will ensure that prices at any given moment more fully reflect underlying realities. And the SEC's database is only one of hundreds that will be boosting economic efficiency, in product markets as well as financial markets, in the on-line age. Economic theory reveals that as information becomes available more quickly and cheaply, markets become more efficient. Resources are allocated more rationally, and wealth is maximized. So at the theoretical heights, not to mention the level at which adults get kicks from their latest electronic toys, the databases look like good news.