diff --git a/2024.09.impact.md b/2024.09.impact.md index 3fe63d9..f57c7f5 100644 --- a/2024.09.impact.md +++ b/2024.09.impact.md @@ -68,7 +68,7 @@ A lot of personal experience, second-hand experience, and observations suggest t While getting people excited about your findings is great, you can often have much greater impact by releasing and growing open-source artifacts that channel your ideas to whatever downstream applications are relevant. This is not easy: uploading code files with a README to GitHub is not enough. A good repository will be the “home” of your project, more so than any of the individual papers you’ll publish. -Good open-source research needs two, almost independent qualities to be good. First, it needs to be good research, i.e. novel, timely, well-scoped, accurate, etc. Second, it needs to have clear downstream utility and low friction. This is the important part: people will repeatedly avoid (and will use!) your OSS artifacts for all the "wrong" reasons, all the time. Just to illustrate, your research may be the objective "state of the art", but people will prioritize a lower-friction alternative nine times out of ten. Reversedly, people will often use your tools for reasons that, to you as a graduate student, miss the point, e.g. because they don't fully utilize of your most innovative components. This is not something to resist; it's something to understand and to build on. +Good open-source research needs two, almost independent qualities to be good. First, it needs to be good research, i.e. novel, timely, well-scoped, accurate, etc. Second, it needs to have clear downstream utility and low friction. This is the important part: people will repeatedly avoid (and other people will repeatedly use!) your OSS artifacts for all the "wrong" reasons, all the time. Just to illustrate, your research may be the objective "state of the art", but people will prioritize a lower-friction alternative nine times out of ten. Reversedly, people will often use your tools for reasons that, to you as a graduate student, miss the point, e.g. because they don't fully utilize of your most innovative components. This is not something to resist; it's something to understand and to build on. Given all of this, here is a list of milestones to observe when working on the open-source side of your research releases.