How to Make a Good Research Artifact
Eric Holk
Abstract
Several conferences, including PLDI and OOPSLA, have added an Artifact Evaluation process in recent years. The goal is to improve the repeatability and reusability of computer science research by encouraging authors to submit artifacts such as source code, proof scripts, videos, etc. to peer review and ultimately publication. Authors who submit artifacts of sufficient quality are invited to display a snazzy badge on their paper, and conferences typically recognize particularly good artifacts at the award ceremony. Although we have generations of precedent for what makes a good research paper, the artifact evaluation process is much newer and thus the expectations for a good artifact are not as clearly understood. I had the opportunity to serve on the Artifact Evaluation Committee for PLDI 2014 and also to submit an artifact to OOPSLA 2014 which was one of three to earn the Exceeds All Expectations Award. Drawing from these experiences, I will share some advice for how to prepare high quality artifacts for future conferences.