Now a County Court judge in the same courthouse, Waite carries with him the stark lesson that eyewitnesses can get it wrong, especially when identifying someone of a different race. No matter how confident, their testimony can "create a terrible result," he said.
Nationally, that criticism has intensified as DNA testing again and again trumps human cognition.
The issue is about to get a rarefied airing before the U.S. Supreme Court, which is to hear oral argument Wednesday in its first significant eyewitness-identification case in 34 years.
But change already is on the docket elsewhere.
In September, a 50-member committee studying wrongful convictions in Pennsylvania recommended that lineup procedures be sharpened to avoid mistaken identification. And in a landmark ruling in August, the New Jersey Supreme Court established a new framework for assessing eyewitness identification.
In the case of Camden defendant Larry R. Henderson, who challenged the eyewitness identification that led to his manslaughter conviction, the New Jersey high court said juries should be better instructed on factors that can affect the reliability of such testimony.
"Eyewitness identifications bear directly on guilt or innocence," Chief Justice Stuart Rabner wrote for the unanimous court. "At stake is the very integrity of the criminal justice system and the courts' ability to conduct fair trials."
Juries have long embraced the testimony of dead-certain witnesses. It helped convict three-quarters of the 275 prisoners nationwide who have been exonerated by DNA over the last two decades. Nine Pennsylvania and three New Jersey prisoners were among those cleared, according to the New York-based Innocence Project.
"This is a really important time for the future of eyewitness identifications," said Brandon Garrett, a University of Virginia law professor who has studied the issue.
As Garrett and others have shown, there are many reasons that a witness can get it wrong.
There is human memory itself, in all its frailty. It can be fickle when it comes to different races. It changes over time, and is susceptible to stress. The presence of a gun also has a significant effect, by shifting attention from the criminal to the weapon.
Misidentification also can arise from flawed lineup or photo identification practices. For instance, police might give signals that encourage a witness to select a specific suspect or validate the choice.
The case before the nation's high court focuses on a narrow question: Should a judge suppress eyewitness identification any time it is made under circumstances that tend to implicate a suspect, or only when police are responsible for those circumstances?
The defendant in Perry v. New Hampshire had been identified by a woman who, as she was being interviewed by police, looked out her window and immediately pegged a man standing with officers as the one she had seen break into a car.
The Pennsylvania Supreme Court has not examined the issue of eyewitness identification in 50 years. But for the last four, a large committee assembled by State Sen. Stewart Greenleaf (R., Montgomery) has.
Although its 328-page report considered a wide spectrum of wrongful convictions, it recommended changes in eyewitness identification. For one, the authors said, line-ups should be conducted by police personnel who do not know who the suspect is. Further, witnesses should be told that the perpetrator might not be in the lineup, and that they should not feel compelled to pick someone.
And, according to the report, police should get statements from witnesses about how certain they are of the identification.
Police and prosecutors on the committee objected to those recommendations, saying they would discourage identifications.
However, Montgomery County Court Judge William R. Carpenter, another committee member, said he hoped they eventually will accept that even a system that works well can benefit from fine-tuning.
"The pendulum in the law swings very slowly, and this is one of those times where the pendulum is starting to move," said Carpenter, who as a prosecutor handled the rape case of a man later cleared by DNA.
Pennsylvania will have a lot of catching up to do, said John T. Rago, a Duquesne University law professor who chaired the wrongful conviction committee.
"If I had to give Pennsylvania a grade," Rago said, "I'd have to say we're in the bottom five in the country."
New Jersey has many of the committee's suggested procedures in place. As early as 2001, the state attorney general, who has jurisdiction over all police in New Jersey, issued guidelines on how to conduct eyewitness identification.
Deputy Attorney General Paul Heinzel said recently that the protocols have had a positive impact on the state's justice system. They help police get more reliable identifications, he said, and "ensure to the greatest extent possible we're getting the right guy."
In Bucks County, Waite, now 67, recalled that he found it humorous to be mistakenly fingered as a robber - at first. But it soon occurred to him that nothing was funny.
What if investigators had taken her identification of him seriously? He had been a public defender for only about a year. "I don't know if I could have accounted for my whereabouts on that particular night," Waite said.
In that case, two men had been charged with robbery. Waite represented one, who pleaded guilty. The other opted for a trial. The victim had seen Waite at the preliminary hearing, he thinks, and remembered him from there, not the crime scene.
Waite said he looked nothing like the defendant, who was probably a foot taller.
Juries often think the word of an eyewitness is the best evidence, he said, but "many times, it's the worst."
Contact staff writer Emilie Lounsberry at firstname.lastname@example.org.