In the months after the summer of 1974, detectives in King County, Washington, wrote down the traits of an unknown man and had names from their files punched onto computer cards. It was a modest act, the kind of clerical labor a later generation would never see, and it produced a list of twenty-five. The man who would be convicted of murder in Florida was seventh on it. The detectives had not acted on the list. The tool had done what it was built to do, and the investigators were not yet in a position to use the answer.
That small story, with its gap between a result and a decision, runs through the first decade in which computers entered murder inquiries in the United States, Britain and Canada. The machines were not clever. They stored, they sorted, and in a few cases they cross-matched. What they could not do was supply the judgment that tells a detective which of ten thousand names deserves a knock on the door. Each early system below failed, or half failed, in a different way, and each failure was written up afterward by someone with a reason to be exact. This is a history of the tool, case by case, and of whether it found the man.
The inquiry began with a public park near Seattle. On July 14, 1974, two young women, Janice Ott and Denise Naslund, were taken from Lake Sammamish State Park on a crowded summer Sunday. Bundy later told investigators that he was responsible, and the deaths are attributed to him; he was never tried for them. What mattered for the tool was what the witnesses supplied: a first name, Ted, a Volkswagen, and a description that put the man in the Seattle area.
Robert Keppel, then a rookie in the King County Sheriff’s Office homicide unit, received the case within days of closing his first murder. The public responded to a published sketch with a flood of calls, at times as many as two hundred a day, by Keppel’s later account. Among the thousands of tips, five named Theodore Bundy, a law student with no criminal record. He did not look like what detectives expected, and the volume of calls about young men named Ted worked in his favor. A person who is one of five among thousands is invisible to a human reader. That is the problem a machine was meant to solve.
Keppel and his colleagues listed the traits the unknown man seemed to have, including the name Ted, a Volkswagen Beetle and ties to the Seattle area, and had names from their files keypunched onto cards and cross-matched. The run took more than a month. It cut the field to about a hundred men, then to twenty-five. Bundy was seventh. The detectives had not acted on the list when, on August 16, 1975, Bundy was arrested in Granger, Utah. Salt Lake County detectives learned what the arrest report might mean, a call went to Washington, and the Washington side confirmed that his name was already on the King County list of twenty-five. Keppel remembered the moment as euphoric, and wrote that the computer result gave him a shudder of excitement.
The list had not caught him. The Utah arrest had. But the list changed what the arrest meant. A name that would have been one of thousands in a drawer was already marked, and detectives from Washington, Utah and Colorado met in Aspen to pool what they had. On October 2, 1975, a survivor picked him out of a lineup in Utah. He was tried without a jury and was convicted of aggravated kidnapping on March 1, 1976, receiving a sentence of one to fifteen years. It was the only conviction any state other than Florida ever obtained against him. He was charged in Colorado with the murder of Caryn Campbell and never tried.
He was arrested in Pensacola, Florida, on February 15, 1978. A Florida jury convicted him of three murders and he was sentenced to death in July 1979 and again in February 1980. On January 23, 1989, the night before his execution, he sat with an FBI agent and agreed that the number of women and girls he had killed was about thirty. Florida executed him at the state prison in Starke on January 24, 1989. He had been convicted of three. Utah announced in April 2026 that DNA had confirmed one more name, a 1974 case.
The Seattle experiment placed the man on paper and still did not reach him. A list of twenty-five names kept in one county did nothing for investigators in Utah or Colorado until a chance arrest connected them. That is a limit of any single-jurisdiction index: a record in one place has no way of meeting a question asked in another.
The Hillside killings, ten deaths of women and girls between October 1977 and February 1978, produced one of the largest task forces the city had assembled: more than 160 officers. The public telephoned in about 10,000 tips. Roger Boren, writing about the trial in 2000, said that the task force kept them on a computer that could hold the information but could not analyze it.
The difference between holding and analyzing is the whole story of the Los Angeles system. A file that returns a record when someone already knows the name is a filing cabinet with electricity. What the investigation needed was something that would notice a name reported more than once by different callers on different days. Kenneth Bianchi’s name was phoned in three separate times before his arrest, according to Boren, and the system never connected the calls.
The break came from Washington State. Bianchi was arrested in Bellingham on January 12, 1979, in connection with the deaths of two college students there, and he pleaded guilty in Washington and, in October 1979, in Los Angeles to five of the murders there. As part of his plea bargain he named his cousin Angelo Buono Jr., a Glendale auto upholsterer, as his partner. Buono was arrested in Glendale on October 19, 1979, and charged with ten counts of murder.
The trial was the longest criminal trial in United States history to that point, by the accounts of the judge, the prosecutors and the wire services. It ran from November 1981 to November 1983, with about 400 witnesses and roughly 1,800 exhibits. Before it began, the Los Angeles County district attorney’s office tried to drop every murder count because it did not trust Bianchi as a witness. Judge Ronald M. George, later the chief justice of California, refused, the state attorney general took over, and a jury convicted Buono of nine of the ten murders. It chose life without parole over death, and Judge George imposed nine concurrent terms on January 9, 1984. Buono died of heart disease at Calipatria State Prison on September 21, 2002.
Nothing in that record shows a computer catching anyone. The case turned on a confession in another state. The Los Angeles file is useful for what it fixes about the technology: by the end of the 1970s a task force could store ten thousand tips electronically, and the stored tips were no more helpful than the paper ones had been. Three calls about one man went unlinked.
The Yorkshire inquiry used paper, and paper is the clearest picture of the problem a computer was supposed to cure. Peter Sutcliffe, a married truck driver from Bradford, killed 13 women and attempted to kill seven more in West Yorkshire and Manchester between July 1975 and November 1980. During those years West Yorkshire Police interviewed him nine times.
The information was kept on a card index in the major incident room at Millgarth police station in Leeds. The index was overloaded and badly cross-referenced. Sir Lawrence Byford’s review, written for the Home Office, concluded that the incident room had become ineffective, with a growing mass of unprocessed material. Officers sent to interview Sutcliffe were often not properly briefed, so evidence that pointed toward him did not reach the people who needed it. He had a gap in his front teeth and size seven feet, features that matched what was known about the killer. The matches were not made.
The inquiry was also steered by a hoax. Letters and a tape from a man claiming responsibility, posted from Sunderland, were given what the government later called excessive credence. In July 1979, at the fifth interview, two detectives, Andrew Laptew and Graham Greenwood, were suspicious of Sutcliffe and said so in a report. It was filed away because his voice and handwriting did not match the tape. Laptew, a junior detective from Bradford, felt the similarities were too many for coincidence and raised his concerns with Detective Superintendent Dick Holland, who made him feel, Laptew said later, so small that he "could have crawled under the crack in the door."
Police arrested Sutcliffe in Sheffield on January 2, 1981, and he confessed two days later. He pleaded guilty to seven counts of attempted murder on April 29, 1981, and a jury convicted him of 13 murders on May 22, 1981, after rejecting his claim of diminished responsibility. Mr. Justice Boreham passed 20 concurrent life sentences. He was moved to Broadmoor Hospital in 1984 after a diagnosis of paranoid schizophrenia, and in July 2010 the High Court imposed a whole life order. He died on November 13, 2020.
The government gave its conclusions on the inquiry to Parliament on January 19, 1982. The police had made "major errors of judgment." Excessive credence had been given to the letters and tape. Had those errors not been made, Sutcliffe would have been identified as a prime suspect sooner. Byford also found that most officers had worked diligently. Among the remedies announced were better training for senior investigators and incident room staff, clearer command for crimes that crossed force boundaries, advisory teams of experienced detectives and forensic scientists, and greater use of computers in major inquiries. The full report stayed confidential for almost 25 years. The Home Office released it, in part, on June 1, 2006. It said Sutcliffe "was probably responsible for many attacks on unaccompanied women, which he has not yet admitted."
That was the official verdict on the paper era. The index had not failed because it was paper. It had failed because no one could see the whole of it at once, and a computer had been proposed as the cure; the full review is a subject of its own.
The inquiry into Robert Black ran on a different principle. It did not search for a pattern in tips; it built a pattern out of the paperwork of a working life. Black delivered posters for a London firm to towns across Britain and to Northern Ireland. He was convicted of the kidnapping and murder of four girls taken between 1981 and 1986: Jennifer Cardy, 9, in Northern Ireland; Susan Maxwell, 11, on the Scottish border; Caroline Hogg, 5, in Edinburgh; and Sarah Harper, 10, in Leeds.
The earlier killings produced no suspect for years. Black was arrested in Stow, in the Scottish Borders, on July 14, 1990, in connection with the abduction of a six-year-old girl. After that, officers from five forces under Hector Clark went through about half a million fuel receipts, expense records and delivery schedules from his employer, The Scotsman reported, and those records placed him near the places where the girls were taken. Detectives spent four years matching the work records to the three murders. There was no confession. The inquiry gathered 189,000 witness statements over the years. Clark, who had led the joint Maxwell and Hogg inquiry as an assistant chief constable of Northumbria Police, was later credited with pioneering the use of new computer systems to manage serious crime cases.
The records cut both ways. Andrew Watt, the former chief superintendent who led the national investigation after 1990, said in 2011 that the file on Black’s teenage sexual offense in the old Argyll and Bute police area had been lost when police and courts were reorganized in the 1960s and 1970s. When officers checked everyone in Scotland convicted of sexual or similar offenses, Black appeared to have no record at all. A system can be only as complete as what is put into it, and no index, paper or electronic, could look for a conviction that had been lost.
Black was convicted at Newcastle upon Tyne Crown Court in 1994 of the murders of the three girls, and received life sentences with a minimum term of 35 years. In 2011 an Armagh jury convicted him of the kidnapping and murder of Jennifer Cardy; Mr. Justice Weatherup set a minimum term of 25 years on December 8, 2011, and the Court of Appeal in Northern Ireland dismissed his appeal on June 27, 2013. He died at HMP Maghaberry on January 12, 2016. Devon and Cornwall Police said that year that he would in all probability have been charged in the 1978 case of Genette Tate; he never was.
The Black inquiry matters to this history because its success depended on a record that already existed, in a company’s files, and on the willingness of five police forces to read it for four years. The computer was a supporting figure. The hard work was the matching.
The Green River killings raise the opposite question. By the time of Gary Ridgway’s arrest on November 30, 2001, the most useful record in the inquiry was not an index but a saliva sample taken from him in 1987. Technology available then could not do anything with it. Fourteen years later, laboratory methods had improved enough that DNA from the sample was matched to evidence connected to three of the 1982 victims, and he was arrested in Renton, Washington. He pleaded guilty on November 5, 2003, to 48 counts of aggravated first-degree murder, and on December 18, 2003, was sentenced to 48 consecutive terms of life without release. A 49th count followed in 2011.
This is a different kind of tool from the card index, and the DNA databases that grew from it, along with the FBI’s national violent-crime linking program, are a subject of their own. The sample matters here only for what it implies. Some of the earliest recorded evidence in these cases was valuable not because someone searched it but because it was kept, and the machine that could read it did not yet exist.
Paul Bernardo’s case is where an official inquiry named the problem in the language of computing and then built the cure. Metropolitan Toronto Police formed a task force on November 17, 1988, to look for a man responsible for a series of sexual assaults in Scarborough, and with help from the FBI it produced an offender profile, according to The Walrus. In May 1990 one woman’s description led to a composite drawing that was widely published. The Canadian Encyclopedia says about 16,000 responses came in, and three said the drawing looked like Bernardo. The Walrus puts the total number of tips at about 41,000.
Detectives questioned Bernardo and, as a routine step, took samples for DNA testing. Justice Archie Campbell, who reviewed the investigation, dates the interview and the blood sample to November 1990. Bernardo was polite and cooperative, and investigators judged him an unlikely suspect, the Canadian Encyclopedia says. His sample joined those of more than 200 other suspects waiting at the Centre of Forensic Sciences, The Walrus reported. The Canadian Encyclopedia says the laboratory, which was receiving samples from across Ontario, had very limited staff at the time.
Bernardo was arrested on February 17, 1993, after the laboratory matched his DNA to the Scarborough assaults in early February. Between the sample and the match, two girls from Ontario, Leslie Mahaffy, 14, of Burlington, and Kristen French, 15, of St. Catharines, had been killed, in June 1991 and April 1992. A Toronto jury convicted him on September 1, 1995, of the first-degree murder of both girls, with other charges, on all nine counts he faced. He was declared a dangerous offender on November 3, 1995, which makes his life sentence indeterminate.
Campbell’s review reported in 1996. In 1997 Ontario announced it would carry out the recommendations and set aside 25 million dollars over five years. A ViCLAS centre, for linking violent crimes, opened in Ontario in December 1996, and reporting to it was made mandatory under the Police Services Act in February 1997. The province set up a serial and predator crime unit, improved forensic laboratory systems and trained major case investigators. The Ontario Major Case Management System was introduced in 2000, and since January 1, 2005, a regulation under the Police Services Act has required every police service in the province to use it. In 2012 British Columbia’s Missing Women Commission of Inquiry quoted Campbell’s findings on systemic failure as still relevant.
Bernardo’s later record is parole hearings. He was refused parole on October 17, 2018, June 22, 2021 and November 26, 2024. A hearing set for October 20, 2026 was reported on September 21, 2026 to have been postponed to January 2027.
Put the six inquiries side by side and the machine’s job description is plain enough. In Seattle it cross-matched traits against names and produced a short list that mattered after the fact. In Los Angeles it stored what the public gave and linked nothing. In Leeds the same task was done on cards by people drowning in them. In the Black inquiry, receipts and delivery schedules became the case, once someone had a name to check them against. In King County a sample was kept until a technique caught up with it. In Toronto the sample was kept too, but the queue it sat in was the failure.
None of them found the man by itself. Bundy was arrested in Granger, Utah, on an unrelated charge; Buono was named by Bianchi; Sutcliffe was caught in Sheffield; Black was caught in Stow; Ridgway was arrested after a laboratory result that the 1987 task force could not have read; Bernardo was matched by a laboratory. The computer’s contribution, where there was one, was to make a name visible after another event had put it in play, or to make a delay plain enough to be counted. The government announced, after Byford, that computers would be used more in major inquiries. Ontario, carrying out Campbell’s recommendations a generation later, opened a centre in which one violent offense could be compared with another.
• July 14, 1974: two women taken from Lake Sammamish State Park, Washington; deaths later attributed to Ted Bundy.
• Late 1974 to 1975: King County detectives keypunch names onto computer cards and cross-match them; Bundy is seventh of 25.
• July 1975 to November 1980: Peter Sutcliffe’s murders and attempted murders; nine police interviews.
• August 16, 1975: Bundy arrested in Granger, Utah.
• March 1, 1976: Bundy convicted of aggravated kidnapping in Utah; sentenced to one to fifteen years.
• October 1977 to February 1978: ten Hillside killings in Los Angeles; task force of more than 160 officers, about 10,000 tips.
• February 15, 1978: Bundy arrested in Pensacola, Florida.
• January 12, 1979: Kenneth Bianchi arrested in Bellingham, Washington.
• October 19, 1979: Angelo Buono Jr. arrested in Glendale, California.
• January 2, 1981: Sutcliffe arrested in Sheffield.
• May 22, 1981: Sutcliffe convicted of 13 murders.
• November 1981 to November 1983: Buono trial; nine murder convictions; life without parole imposed January 9, 1984.
• January 19, 1982: government statement on the Byford review; greater use of computers announced.
• November 17, 1988: Metropolitan Toronto Police task force formed.
• January 24, 1989: Bundy executed.
• July 14, 1990: Robert Black arrested in Stow, Scottish Borders.
• November 1990: Paul Bernardo’s blood sample taken; tested in early 1993.
• February 17, 1993: Bernardo arrested.
• 1994: Black convicted at Newcastle upon Tyne.
• September 1, 1995: Bernardo convicted of two first-degree murders with other charges.
• December 1996 to February 1997: Ontario ViCLAS centre opens; reporting made mandatory.
• 2000 and 2005: Ontario Major Case Management System introduced, then required by regulation.
• November 30, 2001: Gary Ridgway arrested in Renton, Washington; November 5, 2003: guilty pleas to 48 counts.
• December 8, 2011: Black sentenced for the murder of Jennifer Cardy.
• April 2026: Utah announces DNA confirmation of one more Bundy case.