Open Tatianabellagio opened 2 years ago
FlukeAndFeather
commented
2 years ago scientific projects nowadays include not only text as the output, but code and data
In theory the methods section of a paper should describe how the code works, but that's often not enough. Can you think of hypothetical examples of missing steps or ambiguities that might arise?
elhazen
commented
2 years ago A sample with a code library can be valuable to getting the code working - but without the full raw data, it would be hard to reproduce. Do you think the lead author would be against sharing the data publicly or is it simply a lack of funding / time to get the data & code in order?
What version of the prerequisites do you have? This should include RStudio, R, tidyverse, devtools, and rrtools. R version 4.1.1 (2021-08-10) tidyverse:‘1.3.1’ devtools:‘2.4.2’ rrtools:‘0.1.5’ In your words, how would you briefly describe a research compendium? A research compendium is an alternative to just traditional text outputs in research. This is motivated by the fact that scientific projects nowadays include not only text as the output, but code and data. This would enable research projects to be more reproducible.
Consider a paper you contributed to (whether you were on the author list or not). Were the data and code published with the paper? If so, how? Say another scientist wants to reproduce your results. What challenges do you think they would run into?
In the papers I have collaborated on, a sample of the data is publicly available. Not all of it, but it can be provided upon request I guess. Is true that the scripts for the analysis are available but not the version of the programs utilized for them. So, that is a problem that future users could face.