Skip to Content

Reproducible Research

Reproducible Research 不是拿来主义,一味降低难度反而会培养伸手党,currently we assume user 受过高等教育。

How can my paper stay alive, rather than dead on day of publishing? Here alive means others spend little effort while enjoy substantial benefit in using it.

1 readme

We assume user would read before using the software (I prefer “Installation” title).

I won’t explain everything, for example, user is supposed to know install.packages("blogdown") should be executed in R terminal. Luckily, the second result when you Google install.packages leads to a detailed tutorial with screenshots.

2 automatic

Another way is to automatically install dependencies when user run the software. Actually I haven’t come out a fair solution yet.

Convenient as it seems, someone heats changing their runtime environment.

Currently, I assume user would run my workflow step by step, so I just put the installation code on the top. The code is configured to only run in CI environment such as Travis (CONTINUOUS_INTEGRATION=true), as I don’t want to waste time on my computer (install_github() still runs even if the packages is already installed).