Software that can detect and filter e-mail spam incorporates a library of known spam characteristics, learns from user input and recognizes patterns. It works well.The City of Mesa, Ariz., has found that the same process can be applied to infrastructure maintenance. In theory, if software can learn to predict spam, it can learn to predict infrastructure failures. This is the reasoning behind the city’s Sewer Cataloging, Retrieval and Prioritization System (SCRAPS).The Water Resources Department incorporated SCRAPS to help optimize its system rehabilitation and maintenance program. The results to date have been stunning. Hebi Li, the civil engineer in charge of
