Bob Bemer has an idea he thinks could solve the year 2000 problem
that threatens to wreak havoc on mainframe computers after Dec. 31,
Lots of other people are touting solutions, but the 77-year-old
programmer's concept is attracting some interest because he was in
the thick of things when the situation was created.
In the late 1950s, Mr. Bemer developed a key component of COBOL,
which still ranks among the leading computer languages. Much of the
year 2000 problem is embedded in COBOL programs.
"Can you imagine how personally involved I am with this thing,
seeing how they messed it up?" he said in an interview in his Dallas
The problem arose simply enough. Forty years ago, it seemed perfectly
reasonable to use two-digit shorthand for a year, treating 1958 as
58, for example.
But much of this code is still in use today, and when the year
2000 is treated as 00, which could be read as 1900, many experts predict
a myriad of crises in computers used by government agencies and businesses.
Mr. Bemer says there was no good reason to use two digits. "Everybody
makes excuses, saying memory was too expensive, but they're rationalizing,
" he said. "I never used two digits myself."
Mr. Bemer is known for more than 15 critical innovations from the
early days of computing.
He created the "escape sequence" behind the "esc" key on computers.
He is known as the father of ASCII text, making it a worldwide technology
standard. (It's on his vanity license plate.)
Mr. Bemer wrote about the year 2000 problem in a 1979 article for
computer programmers, but after retiring three years later, he dropped
Programmers could have used more time for the job.
Fixing the problem involves tedious work, fixing code line by line,
probably more than programmers can complete in the next 29 months.
Automated tools exist, but they require a lot of handholding by
Mr. Bemer is proposing a far more automated process. Its promise
is that it comes at the most fundamental level at which computers
operate, in the ones and zeros of object code.
Few of today's programmers even know object code, which lies underneath
the languages they work with.
"It's primitive stuff," Mr. Bemer said. But it's a realm in which
he is very comfortable.
And if his idea works, it could prevent the need for a separate
solution for every program.
"We'll run 40 times as fast as anyone else," Mr. Bemer said.
Still, a weakness with any automated method, Mr. Bemer said, is
that individual programmers have their own quirks. Manual intervention
will still be required in some cases.
Fixing the year 2000 problem is a huge job, and Mr. Bemer has tried
to sell the major computer outsourcing firms on his solution.
Plano-based Electronic Data Systems Corp. said it evaluated his
concept but wasn't interested.
"If he does come up with a product, we'll evaluate it then," an
EDS spokesman said.
Mark Ramirez, vice president of a Perot Systems Corp. venture working
on the year 2000 problem, said that in theory, Mr. Bemer's ideas make
"But there are holes in the ideas, and a tremendous amount of time
and energy still needs to be spent going over the concept before it
can be delivered to the market," Mr. Ramirez said.
Mr. Bemer said he has been in tough programming jams before.
"I have every confidence I can make this fly," he said.
PHOTO(S): (DMN: David Woo) Bob Bemer, known as the father of ASCII,
is working on solving the year 2000 problem. Mr. Bemer here shows off
his license plate at his Dallas office. ASCII stands for American
Standard Code for Information Interchange.
© 1997 The Dallas Morning News All Rights Reserved
Alan Goldstein / Staff Writer of The Dallas Morning News,
Pioneer could show the way to year 2000
Programmer takes base-level approach to solving potential computer crisis., 07-28-1997, pp 3D.