Fidelity Investments is a huge mutual fund "family." They began looking at their y2k problem in 1996. They thought they had 50 million lines of code to deal with. Today, they know it's 100 million.
The industry's deadline for testing and full compliance is first quarter, 1999 -- one year away.
This is the assessment of the person in charge of the project: "There is some scope creep. The more you start to look at the problem, the more it grows," she says.
The complexity of the task is revealed in the story in PC WEEK (March 9).
* * * * * * *
Fidelity Investments Co. found it had to rely on its own recipe to cook up a solution to the year 2000 problem.
The group charged with tackling the Boston-based firm's year 2000 conversion in 1996 realized there weren't any commercially available tools in the client/server arena to get the job done -- so they developed their own.
"We had to start from scratch and create the tools to solve the problem," reports Julita Lisowski, a system manager at the Fidelity Technology Center, the group within Fidelity responsible for the majority of the year 2000 central system conversions. . . .
The planning, which began in early 1996, started with establishing the "three-wave approach" to converting Fidelity's threatened systems.
The first wave, started in October 1996 and due to be completed in third quarter of 1998, focuses on regression testing and recoding of all individual applications.
The second wave, enterprise testing, is the most complex, Lisowski says. Started at the end of 1997 and slated for completion in the first quarter of 1999, it consists of technology center staff testing the integration of all the different applications and internal systems to make sure everything will work together. This is the most challenging aspect, according to Lisowski, because officials are trying to re-create the enterprise on a small scale.
The third and final wave of the project concerns an industrywide mandate to begin linking and testing with other companies by the second quarter of 1999.
Estimating the total scope of Fidelity's year 2000 conversion is tricky, says Lisowski, because the original estimate of about 50 million lines of code to be converted has slowly risen to a more realistic number of 100 million lines.
"There is some scope creep. The more you start to look at the problem, the more it grows," she says.