search

atrisovic / dataverse-r-study

Data and code for a large-scale study on research code quality and execution at Harvard Dataverse.
MIT License
17 stars 2 forks source link

Congratulations to the article - questions #14

Open nuest opened 2 years ago

nuest commented 2 years ago

Excellent work :smile:

Comments and questions, I hope you don't mind me misusing the repo for that. Happy to take the conversation elsewhere if you prefer (email?).

  1. The installation of missing packages could also fail when you use an old R version today but CRAN does not have the package for that version, or you use a new R version and R only has it working for an older release. Did you consider using MRAN checkpoints matching the R version release window?
  2. Have you considered using R package that can parse R files for loaded packages? automagic and similar stuff?
  3. Why Python code to analyse R code - personal preference, or did you find that useful somehow?

(Might add more questions as I digest... again, cool work!)

nuest commented 2 years ago

Just found you on Twitter - happy to chat there, too!

atrisovic commented 2 years ago

Hi @nuest,

Thank you! And thank you for opening the issue! These are all good questions.

  1. So, for the first one, no we didn't use MRAN checkpoint and yeah, there were conflicts when installing the older libraries. We see this as both limitations of the analysis, but also limitations of R itself.
  2. Yeah, I considered using goodpractice and lintr (is that what you mean?), but couldn't add them seamleslly into the workflow in this analysis.
  3. Yeah, personal preference and my Python skills are better than R :))

I am happy to follow up - here or over the email also!