New RNG in R 3.6.0 affects order of authors - Githubissues

eric-pedersen / mixed-effect-gams

Repo for tutorial/paper on mixed-effect GAMs

MIT License

122 stars 32 forks source link

New RNG in R 3.6.0 affects order of authors #55

Closed bastistician closed 5 years ago

bastistician commented 5 years ago

The order of authors 2 to 4 has been derived in #28 and is also stated in the manuscript as: https://github.com/eric-pedersen/mixed-effect-gams/blob/aa8dc0754674125ca3c135f1c579abae9a6ef47b/paper_sections/full_document.Rmd#L111-L112

Note that running this code in new R 3.6.0 gives

[1] "Ross"    "Simpson" "Miller"

instead of

[1] "Miller"  "Simpson" "Ross"

which can only be reproduced after setting RNGversion("3.5.0").

So reporting the seed is no longer sufficient for reproducibility of this part of the manuscript. :wink:

eric-pedersen commented 5 years ago

Ha, of course it did... Thanks for the heads up! That's really useful to know, and likely affects the reproducibility of the other sections as well (I'm okay without the authorship list being 100% reproducible. :smile: ); I'll add a fix to check for the version of R and set the RNG back to legacy behaviour if R version > 3.6.