Previous StoryNext StoryTable of Contents

PreText Magazine is produced by PreText, Inc., a digital media studio specializing in content, design, and technology for the Web. For more information visit the PreText Web site.

Outsmarting Y2K

"Sometimes I regret creating Cobol," says Bob Bemer of the still-dominant machine language for which he built a key component in the 1950s. "It allowed lots of people that are less competent and responsible than they should be to get into the computer field."

Bemer says that programmers' lack of knowledge of object code--the code based on ones and zeros that computers operate on at the basic level (and upon which COBOL is built)--is at the root of the impending year 2000 computer crisis, commonly known in computing circles as the Y2K problem. Forty years ago no one worried about using two-digit shorthand: '98' to represent 1998, for example. "It was too expensive to create the extra memory; there were lots of excuses. And then as technology advanced, people got more interested in being on the 'cutting-edge' than working on the stuff that lies underneath the languages they work with now."

"Can you imagine how personally involved I am with this thing, seeing how they messed it up?" he exclaimed during an interview with the Dallas Morning News in July. But now the 77-year old computing pioneer (who also had a hand in creating the ASCII universal character naming standard and the technology behind the ESC or "escape" sequence that sits in the upper left-hand corner of your computer keypad) thinks he might have the antidote.

Bemer claims to have developed an automatic fix for the most expensive problem now facing the computer industry. Instead of the line-by-line fixes proposed by many in the trade (a solution Bemer thinks would take more than the short time left until 2000), his Vertex 2000 software goes to the heart of the problem: the object code.

Once installed in a company's computer system, Vertex extracts all the arithmetic instructions involving dates. Special subroutines based, for example, on the fact that dates are never multiplied or divided, identify the date strings. Each time a date instruction is found, it's uploaded to a "virtual machine" or holding pen where a collection of extra bytes are attached creating additional digits--what Bemer calls BIGITS--that represent the century. Once the year-date has been fixed, it's sent back to the original application. Vertex also includes subroutines for defining new dates entered into databases as BIGITS, so that neither existing databases nor input forms have to be changed. Because Vertex modifies the object code, not the source code, the modified data can still run the old programs. Industry leaders, even the U.S. government, are taking note of Bemer's claims.

Speed is possibly the biggest selling point of Bemer's invention. Vertex cuts testing time down so that instead of having programmers test each and every line of code, a painstaking process that could take months, companies can automatically pull operations that need rehabilitation. This gives them more time to find and fix glitches unique to their systems and get back into production fast. Dramatically, Bemer believes the entire process should take a maximum of 15 days. Bemer admits that Y2K is an intensely complex problem, but he thinks he just may have the silver bullet.

    --by Susan Dumett