IRC channel logs
2022-11-29.log
back to list of logs
<rekado>ACTION updates CRAN packages yet again… <old>Quick question unlerated to Guix here, but if you were to cite a code repo in a paper/thesis, e.g. gitlab, github, would you use a URL to the repo or one to software heritage if available? <zimoun>« Reproducibility efforts as a teaching tool: A pilot study » <zimoun>Quoting: «Regarding difficulties, all students reported similar experiences. There were 2 recurrent observations. First, nearly all groups reported that no code had been made available by the authors or their assigned article.» <zimoun>«Second, all groups reported the absence of a detailed explanation of how each analytical method was applied. The groups overcame this difficulty by relying on the documentation and user guides available for each package and by trying out different parameter combinations to “guess” the ones used by the articles’ authors.» <zimoun>«Furthermore, in cases where some code was provided, it lacked documentation and rerunning it generated some differences from the published results, which could lead to different scientific conclusions. We believe that this trial-and-error process was one factor that made the projects so time-consuming and resulted in only 11 out of 28 (39%) figure panels conveying the same information.» <civodul>zimoun: thanks for the link, i hadn't seen it <civodul>not really surprised by the findings :-) <zimoun>heh! I would be interested by similar results about student who knows how to program (CS student or “engineering”). <rekado>almost done with the CRAN updates <rekado>DT took more time than all other packages combined <rekado>ACTION builds pigx now as a test case <civodul>i guess all this work makes Guix the best distro for anyone who uses R! <zimoun>About emacs-ess, it is also in my TODO list. But… more items are coming than I am able to push other out. :-)