I have the impression that the Clean team implements it's I/O libraries for every OS again and again instead of simply interfacing with an already existing library that provides OS abstraction. Am I right? Do they try to reinvent the wheel? Brian Woods