The traditional history of the dot-com bubble has been told many times: Too many companies rushed into the market in defiance of all known business fundamentals, and when the crash came, all but a tiny fraction of them just as quickly imploded and went away.

That received wisdom, though, is now getting a going-over by economists, business historians and others, some of whom are coming to new conclusions about what precisely went wrong during the bubble years, normally dated from the Netscape IPO in August 1995 to March 2000, when Nasdaq peaked at above 5100.

A recent paper suggests that rather than having too many entrants, the period of the Web bubble may have had too few; at least, too few of the right kind. And while most people recall the colossal flops of the period (Webvan, pets.com, etoys and the rest) the survival rates of the era's companies turns out to be on a par, if not slightly higher, than those in several other major industries in their formative years.

The paper is being published in a coming issue of the Journal of Financial Economics. As noteworthy as the findings are, even more interesting is the process that led to them. The work is an outgrowth of the Business Plan Archive at the University of Maryland. Its goal is to become a kind of Smithsonian Institution of the Internet bubble, saving for posterity every business plan, PowerPoint presentation and venture-capital term sheet -- the more frothy and half-baked, the better -- that it can get its hands on.

David A. Kirsch, a professor of management at the university's business school and one of the authors of the study, said it relied on a thorough examination of one particular treasure trove at the archive: every business plan, roughly 1,100 in all, submitted during the period to an East Coast venture firm.