|Ottawa Citizen Online|
Monday 5 October 1998
For decades, experts squandered chances to fix year 2000 problem
Extra digits were very expensive in early years; industry then never wanted to face the problemDavid Hayes and Finn Bullers
Knight Ridder Newspapers
Richard Nixon will have been dead for more than five years on Jan. 1, 2000, but the way Bob Bemer tells it, the year 2000 computer bug is another thing we can blame on the former president.
In about 1969, Mr. Bemer was a leading programmer pushing for Mr. Nixon to declare a National Computer Year to raise awareness of the nation's growing reliance on technology. Mr. Bemer's agenda for the year-long program included a push to drop the old two-digit standard used by programmers to identify the year in computer programs.
Even three decades away from the year 2000, the computer pioneer was worried that using only two digits to represent the year eventually would cause problems.
Although Mr. Nixon's science adviser was on the National Computer Year planning team, the president wouldn't sign a proclamation. He didn't believe in computers, said Mr. Bemer, now 77 and president of a software company that is marketing its own year 2000 fix. The two-digit year remained the norm.
OK, so that doesn't make Mr. Nixon the parent of the year 2000 bug -- but it was another opportunity missed.
In fact, scores of opportunities were neglected for decades by many people aware that computers would misread the year once 2000 rolled around.
It was all a matter of time and money, said Ed Yourdon, another early computer programmer who now makes his living consulting and writing about the year 2000 problem.
The "19" was dropped from dates to cut keystrokes, to save room on computer punch cards and to save space in expensive computer memory banks.
"Twenty or 30 years ago it was an engineering trade-off," said Mr. Yourdon, chairman of the Cutter Consortium and co-author of Time Bomb 2000: What the Year 2000 Computer Crisis Means to You.
"To build some of the early computer systems, there were so many tricks, so many trade-offs, to make them work at all," Mr. Yourdon said. "A megabyte of memory was $100,000, compared to $1.98 today."
Mr. Bemer said productivity was an even larger issue. Although some programmers proposed using four digits for the year, their bosses said it was a waste of time to have keypunch operators repeating the same two digits -- 19 -- over and over.
Add to those issues the freewheeling atmosphere in which programmers worked, which makes problem code even harder to find and fix now. Unencumbered by rules, programmers sometimes named areas of programs that contained dates after their girlfriends or favorite sports teams, leaving them buried in thousands of lines of com puter code.
The glitch was passed on from one generation of programmers to the next. So today, businesses and governments around the world are scrambling to fix a problem that threatens everything from the national electrical power grid to some VCRs.
Mr. Yourdon said programmers never worried then about what would happen at the turn of the century because they assumed their work would have been replaced long before then.
"We had the impression that not only would the hardware be thrown out in seven or eight years, the software would be thrown out, too," he said.
But changing software in mainframe computers that handled accounting and other tasks turned out to be too expensive for corporate America. The software stayed around, even as old hardware was replaced.
Even though memory became less expensive in the 1980s and downright cheap in the '90s, no one wanted to tackle the issue of changing dates from two to four digits.
It was difficult to convince corporate executives that spending $3 million to fix something that wouldn't add to the corporate bottom line was a necessary expense, Mr. Yourdon said. "They'd say, "Are you nuts? Somebody who takes my position three or four years from now can worry about it."'
Leon Kappelman, a prof essor at the University of North Texas and author who thinks the United States will face serious problems on Jan. 1, 2000, said the industry might be headed for serious changes if he's right.
If widespread problems or deaths are caused by the year 2000 problem, Mr. Kappelman said, the computer industry is likely to face significant regulation.
Copyright 1999 The Ottawa Citizen