Pushing for Quality Data Design
Industry needs to acknowledge ‘huge blind spot’
BY MARGO McCALL
Scott Ambler hasn’t seen too many places where data and quality intersect. One exception from the real world is the intersection of Data and Quality streets in the city of Sacramento, California. However, even that intersection features a prominent stop sign.
Ambler, the outspoken expert on software process improvements, believes the database community needs to acknowledge its “huge blind spot” and make the fundamental changes need to resolve the major challenges it faces.
“The database community needs to come up to speed on this big whack of stuff it’s chosen to ignore,” he told those attending this year’s SD West conference.
According to a 2002 Data Warehousing Institute report, bad data costs US businesses $611 billion annually. “Where I come from, that’s real money and maybe we should try to do something about it,” said Ambler, IBM Rational practice leader for agile development. “I don’t know how they keep their jobs.”
While the majority of database managers he surveyed acknowledged they have serious production database problems, workable solutions remain elusive. Eighteen percent of the respondents had no strategy; 33 percent were trying not to make the situation worse (a “nonstrategy” in Ambler’s opinion); and another 8 percent had “fantasy” plans to rewrite everything.
According to Ambler, the only viable solution is an evolutionary approach, a strategy embraced by the remaining 33 percent. The behavioral approach he advocates is driven by tests, not models, and reflects the philosophies of agile development. The traditional approach, which features long feedback cycles and early decision-making, is expensive and known to result in serious mistakes.
Ambler urged developers to learn about databases. “The database is one of your layers. You might want to take it seriously,” he said. “It’s time to grow up a bit and get some data skills.” CW (March 2008)