In the pursuit of any goal, the first step is invariably data collection. As put up on the [OpenAI](https://openai.com/blog/special-projects/) blog, writing a program which can write other programs is an incredibly important problem. This dataset collects publicly available information from the [Codechef](http://codechef.com) site's practice section to provide about 1000 problem statements and a little over 1 million solutions in total to these problems in various languages. The ultimate aim is to allow a program to learn program generation in any language to satisfy a given problem statement.