"I remember them in tweed caps, wrapped in blankets on the first-class decks of ocean liners; or strolling along Fifth Avenue on Sunday mornings in their top-coats and hats from Locke; or in Hunterdon County in autumn with Faulkner down for the weekend. They lunched at '21' and dined at Chambord and the Colony. After a Sunday lunch at the Knopfs' in Purchase [N.Y.], the shades were drawn and home movies shown of Alfred [Knopf] and Thomas Mann in lawn chairs beside Lake Constance, flickering like automata as they moved their arms in conversation."

That's the celebrated Jason Epstein, for many years the editorial director of Random House, remembering big-name Jewish publishers as he saw them when he entered the business in the 1950s. Those lions had been innovators in their time, in the 1920s, overturning the dominance of the genteel old-line publishing houses whose tastes "were rooted in the prejudices of the previous century."

Epstein would become an innovator in his own time, beginning the quality paperback revolution with Anchor books in the 1950s, helping to found the New York Review of Books during the New York newspaper strike in 1963. His 2001 memoir, Book Business: Publishing Past Present and Future, is wonderful reading for anyone seriously interested in the trade.

Especially interesting is the account of his attempt in the early 1990s to get The Readers Catalog off the ground. This three-inch thick paperback, a comprehensive inventory with which readers could order books from a toll-free telephone number, was designed to be a "worldwide bookstore without walls." The 2,000-page list of 40,000 titles hit the bookshops just about the same time that the World Wide Web made its appearance. Today its second edition is a very minor part of Barnes and Noble's online bookselling empire.

The technology of publishing moves on, and today the innovators who are pushing out its frontiers are electronic magnates.

Take Michael C. Jensen, proprietor of something called the Social Science Research Network. He is an innovator as different from those who have gone before in scientific publishing as were Knopf, Bennett Cerf and the others who revolutionized trade publishing in the 1920s.

Better known today than SSRN is the National Bureau of Economic Research, which, among its many other activities, performs a similar service -- publishing research to a professional audience, before it has been accepted by a scholarly journal. But SSRN is much bigger, more far-flung and potentially far more significant, because it casts a wider net.

Both publishing enterprises owe their influence to something that might be called "the working paper revolution," a response to an explosion in the volume of published research in the years since the 1970s.

When an economist finishes a paper -- or business school professor, legal scholar, academic accountant or finance economist -- he or she sends it off to a scholarly journal and settles down to a long wait.

The editors of the journal date-stamp the manuscript when it arrives, read it (in turn with all the other manuscripts that arrived that day), then send it off to a pair of "referees" who have agreed to work for them, usually for free, as expert advisers whose job is to report to the editor on the quality of the submission.

Almost inevitably the referees -- who are protected by at least the pretense of anonymity -- will require revisions, for good reasons and sometimes bad (has their own work in the field been insufficiently mentioned?). By the time the paper is accepted and finally published, as much as two or three years can pass. This in a highly-competitive business in which being first with a truly good idea can make all the difference in a career.

No wonder, then, that various intermediaries came forward with a view to shortening the interval before research could be freely shared among colleagues, in anticipation of an eventual acceptance by a journal. For a time, university departments took responsibility for their own members, issuing "preprints" with a quick-and-dirty certification that the paper would meet the standards of the field.

Then the NBER stepped in, providing "working paper" status for submissions by its own researchers -- some 300 of the best and brightest scholars in economics around the world, as hand-picked by the Harvard and Massachusetts Institute of Technology dons who dominate "the Bureau."

And for a time in the early 1980s, NBER "yellow-jackets," as the little yellow-clad pamphlets came to be known, were virtually synonymous with macroeconomics' research frontier. Nothing else mattered very much.

Other fields evolved similar mechanisms -- small groups of people from the very best institutions, meeting each other in seminars and workshops, organizing conferences to discuss their work. Field after field was thus informally organized as a relatively small Major League and a relatively large Everybody Else -- very convenient for journalists, among others.

But that clubbiness was too exclusive -- the sensibility too distinctive -- to maintain pre-eminence for very long, especially in the digital age, with perestroika in the air. And of course meanwhile the Internet had been invented, converted from a small government-owned information pipeline serving mainly defense researchers to broad-gauge information highway available to all.

Jensen remembers getting a call in 1992 from then Clemson University finance professor Wayne Marr. The Internet browser had yet to be invented. With a colleague, John Trimble from Washington State University, Marr was putting together a finance discussion group. Would Jensen lend his name?

Jensen had a history of this sort of thing. For one thing, he was a research star of the first magnitude -- his 1976 paper "Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure" had transformed the study of principal-agent relations and, in the process, provided some of the intellectual underpinnings for the corporate restructurings that shifted into high gear in the 1980s.

For another, he had a history as an entrepreneur. In 1974, Jensen had founded, with Eugene Fama and Robert Merton, the Journal of Financial Economics. That widely-read periodical, which burst upon the fast-changing scene at the moment of its flood, was widely seen as having changed economic publishing, with its high submission fees, paid referees, dramatically shorter turn around times and general transparency.

A discussion list? Sure, said Jensen, before leaving on a month-long business trip. By the time he returned, the business had changed. Now Marr wanted to distribute abstracts of completed but not-yet-published new work in finance. And so the he Financial Economics Network was born.

There were the usual difficulties. Trimble wanted to charge large sums for the service; Marr left eventually to pursue other interests (a similar network devoted to Biblical studies). "I ended up holding the bag for this thing that I never really started, but now I felt responsible for," says Jensen. "I don't like to give up on things. I like them to succeed."

So in came Fama, to give the organization additional clout. The progenitor of the efficient-markets hypothesis, he was a fixture at the Graduate School of Business at the University of Chicago; Jensen was a senior figure at the Harvard Business School. Together they patched up the enterprise. For Jensen, it was an "experiment in organization." It was also a way of returning to where he had begun.

"I served as an apprentice printer in a job shop in Minneapolis, while I was going to high school and [Macalester] college. [He graduated in 1962] I came full circle, back to my roots, in a different technology."

Gradually the Social Science Network Research emerged. The servers were moved from Austin, Texas, to a site near Washington D.C. A chief executive was hired. The National Bureau of Economic Research joined forces in certain key undertakings; its president, Martin Feldstein, agreed to serve as co-director of the Economic Research Network with Jensen.

Today, SSRN is built around a collection of eight tightly-moderated research networks, all operating with the openness and organized skepticism of science, and casting the widest possible net among serious researchers -- the original Financial Economics group, an Economics network, Accounting, Legal Scholarship, Information Systems, Management, Marketing, and Negotiations.

That means that most of the world that was left out before can now participate fully in the construction of economic understanding -- contributing not just abstracts that hint at promising ideas and which contain contact information, but full text documents as well -- including the working papers of every major law school in the world. There are some 60,000 papers in the system, some 40,000 of them full-text versions, with downloads running at the rate of 180,000 a month, and abstract views running at around three times that rate. Some 38,000 researchers around the world are involved.

Moreover, the volume of scholarly papers available through SSRN is soaring, though often it costs money to read them. "We don't charge for the downloading of papers, except when the owner of the papers charges us," says Jensen. "The NBER charges us, so we charge the users. We distribute the papers of over a hundred journals, online, and most of the publishers want to charge. So we charge. Other than that, we give away the database freely." It's divided now into unpublished "working papers," and an "accepted paper" series, which includes recently published work as well.

And who decides what will be accepted? The answer is, the authors themselves. "We apply no filters other than that the paper should be part of the worldwide scholarly discourse in the area," says Jensen. So once in a while, an editor, or I, or somebody else, will send back a paper as inappropriate. But other than that, we've followed the principle, 'Let a thousand flowers bloom.'"

Many of those flowers are doomed to quickly wilt, of course. But the papers that will be deemed great in the future will show up here first too. As the banner says, "Tomorrow's Research Today." You only need to know what you're looking for.

Thus one of the biggest bargains around is the regular email announcement of newly arrived work. For something like $15 a year, you can get an email journal of abstracts that arrives once a week on your desk with all the latest stuff. You can customize it if you like.

You don't have to be a professor to know how valuable such a digest of current commentary can be. By now, nearly everybody must subscribe to several services that more or less define the communities of which they consider themselves to be working members.

I'll mention, for example, one report that I read every day -- Jim Romenesko's authoritative digest of US media industry news, commentary and memos for Poynter Institute. I subscribe to Johnson's Russia List as well. It's an exemplary service that provides headlines and full text stories about Russia from both the English language and the Russian mainstream press, often several times a day.

Once a week or so I get an email list of new NBER working papers. And on those occasions when I take a peek at the SSRN networks, I feel as though I am drinking from a fire hose -- it leaves me feeling bloated, soaking wet, but at least potentially fully informed.

Today, SSRN is on the verge of becoming a major publishing operation. Discussions are ongoing with a wide array of organizations that may join. There are working papers to be posted, conferences papers to be hosted and archived in the e-library, still more journals to be brought under the roof. "We've put millions of dollars into it," says Jensen, who regularly turns down inquiries from commercial publishers about the possibility of a sale.

"Someday, I hope it will return something like the cost f capital, and pay the interest on its debt. But we have created this thing, and got our colleagues around the world to put their babies up there -- because that's what this work is to them, their children -- and now we have to worry that it never falls in the hands of somebody who's going to try to capitalize on it in a short-run way." Ahead lie bold experiments in the new technologies known as collaborative filtering. Might it be possible to do away with referees altogether, or at least dramatically reduce their role? Will editors and publishers be next?

I doubt it. Jensen himself a colorful figure, with his University of Chicago PhD, and the nearly twenty years he taught at the University of Rochester. It was there, with the late William Meckling, that he made his seminal contribution to economic knowledge.

In 1985 he moved to Harvard Business School and completely (and contentiously) revamped its curriculum in organizations and control, before decamping to the Monitor Group consulting firm across the river in Cambridge, where he has been managing director of its organizational strategy practice since 1999.

But more of that some other time. For throughout, he has remained chairman of SSRN -- a figure hardly less interesting in the present day than the great insurgents who tore up book publishing in the 1920s.