-A
Computer science and computer engineering is, for the most part, the study of software. Surprisingly, most of the people who do this research have never written programs that others use ("real software"). The arguments against producing real software are painfully clear: it entails considerable work and long-term commitments. This talk focuses on the less obvious, but equally compelling arguments in favor of producing, documenting, and distributing robust, usable software as a routine part of research. In my experience, real software yields a wide range of benefits, both personal and professional, and should be the norm, rather than the exception.