PALO ALTO, Calif. (CNN/Money) - Three years ago, nobody thought much of history: In the "New Economy," the rules were being re-written.

Now, everyone is looking backward: How long do bear markets last? Is the U.S. in the oughties the same as Japan in the 1990s? What were appropriate market valuations historically, and therefore what should they be now?

Taking history lessons right now is quite appropriate, especially for understanding the broad market. As far as individual companies go, history can be downright scary.

Take Sun Microsystems, the embattled computer server maker whose stock keeps falling. At its news low levels, is Sun a good buy yet?

The oft-quoted and highly quotable research analyst Steven Milunovich at Merrill Lynch conducted an exercise in history Monday by publishing a report on Sun's valuation relative to its historical trading levels. His findings are instructive, even if they understate the woes facing Sun investors.

For example, Milunovich noted that Sun has a P/E of 33, based on his earnings estimate for fiscal-year 2003 of 13 cents per share. He notes that in 1991, Sun had a comparable P/E of as little as 7.

If it traded for that valuation now, the stock would be just 91 cents a share, down from around $4 recently.

Could a company as large as Sun (SUNW: down $0.19 to $4.06, Research, Estimates) really become a penny stock? Sure, but not just because it did so in 1991. That year Sun had revenues of $3.2 billion, compared with revenues of $12.5 billion in fiscal 2002. When Sun was relatively puny investors didn't give it the benefit of the doubt. What Milunovich's noodling shows is that even today it's getting a bit of a pass. If it went down to the ignominious level the Merrill analyst is suggesting, it would be because investors think Sun's strategy of selling proprietary hardware and software is bankrupt.

Milunovich also mines some interesting data in noting that Sun's price-to-trailing-sales ratio of just over one times isn't so low. Indeed, he notes that between 1989 and 1995 Sun traded at or below one times sales for six years.

RECENTLY BY ADAM LASHINSKY       AOL's cookie jar The 'real' options problem Making the 'sell' decision

Frighteningly, he notes that 0.4 to 0.5 times sales marked the bottom multiple for bygone names like DEC, NCR and Unisys. Slapping that valuation on Sun suggests another 50 percent drop in the share price. NCR and Unisys are still around, of course. But they're hardly what Sun Microsystems wanted to be when it grew up. (NCR and Unisys still trade for around half their sales.)

The point of Milunovich's exercise seems to be to urge investors not to get caught in a classic value trap. Just because something's down doesn't make it cheap. Fair enough. The next leg needed from the venerable analyst is a strategic review of whether Sun will re-emerge above its trough valuations and why.

Adam Lashinsky is a senior writer for Fortune magazine. Send e-mail to Adam at adam_lashinsky@timeinc.com.

Sign up to receive The Bottom Line by e-mail.