Hacker News new | past | comments | ask | show | jobs | submit login

> Is there any incentive for reproduction of experiments?

There're several incentives to make your research as obscure and unreproducible as possible. First, you want to keep most of the details private to your lab. Why would you help other labs? So that they can make a discovery before you? No way!

Next, suppose it was easy to re-run your whole data analysis pipeline outside your lab. So there's someone you don't even know who tries it, and finds and tiny bug in your code that invalidates your results. Now you have to publish a correction, or even retract. You don't want that happening ever. So you publish some undocumented and unrunnable code so that nobody would even bother reproducing.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: