
Snags, Traps, and Black Holes ◾ 231
A few years back I was quality reviewer on a data warehouse project that was being deployed,
and although the software worked according to specication, the inputs and outputs of the
tool were gibberish. As the technical team dissected the data ows, they found that there were
data conicts that were irreconcilable. Not only were there duplicate and disparate source
data feeds, but they uncovered data transformation algorithms that actually corrupted good
data. In peeling back the onion, data accuracy was at the heart of the issue. With all the data
conicts, it was estimated that the data feeding the data ...