As sniper suspects John Lee Malvo and John Allen Muhammad were making their way into courthouses in Virginia last week, Bernard Webster was making his way out of one in Maryland.

Webster, convicted of rape in 1983, had spent nearly 20 years in the Maryland Correctional Institution at Hagerstown before DNA testing revealed that he had not committed the crime.

Here, then, is yet another cautionary tale about our imperfect judicial system, and how the more certain we are of a suspect's guilt following an arrest, the more careful we ought to be about ensuring a fair trial.

The evidence against Webster certainly appeared to be a slam-dunk. Three eyewitnesses, including the rape victim, had identified him as the assailant. Moreover, a chemist with the Baltimore County police crime lab had testified that DNA samples taken from the victim's body matched Webster's.

So what if Webster had not been provided with an attorney during the lineup, as he was entitled to by law. Ruling on Webster's request for a new trial, the Maryland Court of Appeals found that failure to provide counsel "was not an error of constitutional dimensions" and let his conviction stand.

In the case of Malvo and Muhammad, so much pretrial evidence against them is now being made public -- much of it obtained by interrogating 17-year-old Malvo without an attorney present -- that the only thing more certain than a prejudiced jury pool is that the pair will be convicted and, barring a plea deal with Malvo, sentenced to death.

But as we replace blind justice with this thinly veiled eagerness to execute, it's worth remembering why more care is required to make our judicial system honest and impartial.

In 1982, Webster, then 19, was in the Baltimore City Jail serving time for a misdemeanor assault when he was charged with rape. He had been picked out of a lineup by the victim and two others who claimed to have seen him in the vicinity of the crime.

The police chemist testified that a Type A and Type B blood group had been found on swabs taken from the victim. Webster is Type A; the victim Type B, and it was the chemist's expert opinion that the rapist was Type A.

Curiously, the chemist's written report noted only a Type AB blood group. Asked by defense attorneys to explain the discrepancy, she replied that her written report had contained a typographical error.

Although Webster produced two witnesses who said they were with him on a basketball court at the time of the rape, he was convicted and sentenced to 30 years in prison. And because he refused to confess to a crime he did not commit, he was given hard time at Hagerstown instead of a more rehabilitative term at the Patuxent Institution.

In March 2001, after having been denied parole five times, Webster took advantage of a relatively new statute that allows DNA testing for convicts. Michelle Nethercott, chief of the Innocence Project of the Office of the Baltimore Public Defender, helped secure three swabs taken from the rape victim that were still on file at the hospital where she had been examined.

The tests, performed by the Serological Research Institute in Richmond, Calif., were conclusive: The DNA was not Webster's. The police chemist, as it turned out, had not made a typo in the written report. Her court testimony was wrong.

And just whom those witnesses actually saw near the scene of the crime, God only knows.

Webster, now 40, was set free Thursday -- penniless and without apology, as if keeping an innocent man locked up for half of his life was no big deal.

What is considered a big deal is that Virginia has "the best range of available penalties," as Attorney General John D. Ashcroft recently put it, meaning that even juveniles can be executed.

And if there ever are questions about the guilt of the condemned, so what? After convicted rapist and murderer Roger Keith Coleman was executed by the commonwealth in 1992, several media organizations and a charity sought permission to conduct DNA tests to help lay to rest the many questions that remained about his guilt. This month, the Virginia Supreme Court denied the requests, ensuring that if the wrong man had been executed, the public would never know.

So we may well consider Muhammad and Malvo as good as dead. But as the Webster case reminds us, it doesn't always matter whether you're guilty.

E-mail: milloyc@washpost.com