Thesis
An #Rstats script written 20+ years ago is much more likely to run today, without modification, than an R script written 7 years ago, especially if the latter used any part of the #tidyverse
@davidbraze that’s probably true, but how important is it? Not being able to change a system makes it much harder to improve. IMO that limits the audience of future users by over weighting current users.
@hadleywickham @astrodino
@brodriguesco
@eamon
@nibsalot
@astrodino
1st, I'm against neither innovation nor the tidyverse. I use tv regularly
But it is important for users, maybe especially new users, to be cautious about what non-core packages they latch on to in their own work
And cognizant of the fact that the farther you get from `installed.packages(priority="high")`, the more risk you must accept with regard to bit-rot and the tech bro (toolmaker?) ethos of "move fast and break things"
@davidbraze @hadleywickham @astrodino @eamon @nibsalot Whether you use noncore libraries or not, there are so many things that can make a 20 old script not reproducible that just not using noncore packages is not a safe guard in any way.
just wrote about this today https://www.brodrigues.co/blog/2023-01-12-repro_r/