Search code, repositories, users, issues, pull requests...
Abstract. The future of scientific papers lies in the tight integration of traditional text with program code that can reproduce results on demand. In this essay, I demonstrate that this integration also enables new reuse mechanisms: In the future, we could import and extend code from other papers just as we cite and extend their ideas – "from Miller1989 import fuzzer".
The difficulty to reproduce, reuse, or extend scientific results has long been a common concern in scientific circles. The problem is the more troubling as computer software is commonly used to evaluate or even embody research results - and while one may assume that computer software should be far easier to copy and extend, most scientific software is not in a shape that others could work with it. Even in computer science, where the majority of all publications describe some software or algorithm, only a small minority of papers actually makes this software available to others.
Many reasons are cited against making software available, mostly the fear and cost of having to support some huge infrastructure. And yes, if you have implemented, say, a symbolic analysis for FORTRAN-77, you will end up with a huge mess of code that will be hard to maintain. On the other hand, there will be a few key contributions your analysis will have made – the key contributions you would like to detail in your paper, possibly with pseudocode. The question is: Aren't these key contributions worthwhile not only to be read, but also to be used and reused?
