The home of an elderly couple in Brooklyn, New York was visited by police “50 or so” times over a span of 8 years. If you’ve had the police come for a visit, you would know that they don’t bearing gifts. When the police come a knocking, they might pound your door in with a battering ram and kill you dogs. So you can only imagine what fright the retired couple had to endure. The numerous police visits where traced to a software glitch in a computer system used by the New York Police to track crime complaints and criminal activity. As it is customary for Associated Press reports, there is very little information or follow up on the nature of the software system in question. The exact details behind the computer error were not given, other to say that the error started in 2002 when the police upgrade from a manual process to an automated computer system.
From working with a variety of computer systems, I know how an error like this can potentially have been introduced. Often times, when working on a new software feature, you have test said features but with fake data. A common practice is to simulate a small portion of the computer system with fake data to mock the environment. In the worst situations, actual test data or test conditions are hard coded in the actual application. If fake test data is embedded in a production system, like that used by the New York Police, their might be certain conditions like a certain date or time or report type that will trigger the test data to percolate to the surface.
Along this lines, I have seen certain feature in a computer system not function correctly because it is installed in a Windows Vista as opposed to Windows Vista, or that on leap years it behaves erraticly, or that if you installed it on the D: drive as opposed to the C: drive you won’t be able to save files, etc.
As we wrap database and computer systems around every piece of personal data, from credit report to no fly lists, it is important to design them in such a way to limit the number potential victims of said systems. For example, if you are a victim of identity theft you will have to go to great lengths to clear your name and credit history because of how these systems are replicated and copied and ultimately considered to never be wrong.