This is a paper written by Raymond R.Panko and Daniel N.Port from the University of Hawaii. It looks at the costs and risks that organisations carry as a result of having thousands (millions) of applications (developed by end users) that (a) keep the business running (b) are out of sight from most management procedures and IT control.
While it has a very strong "science" flavour to it (after it, it is being written by academics), it does a good job of identifying and explaining the issues that plague modern businesses when it comes to the explosion in spreadsheets and other "end user" applications. It avoids jumping to the naive and unrealistic conclusion that businesses should do away with spreadsheets and end user applications, instead, it opts to focus on understanding the problem first, setting the stage for pragmatic solutions.
End user computing (EUC) applications are like dark matter in physics. They are enormous in quantity and importance yet have been largely invisible to corporate IT departments, information systems researchers, and corporate management. EUC applications, especially spreadsheets, are also “dark” in the sense that they pose a number of overlooked risks for organisations, including errors, privacy violations, trade secret extrusions, and compliance violations.