Bela Lugosi's Dad on Nostr: Starting to think that reproducibility in research is a great idea, but probably ...
Starting to think that reproducibility in research is a great idea, but probably rarely obtainable in the real world for real analyses. And maybe that's okay? If you were trying to reproduce a physical experiment from a decade ago, you're going to have to set up the equipment, maybe the equipment is a bit old and dusty and you have to replace some fuses. Nobody said it would be easy and instant. But if you're the original person doing the analysis, you can at least *try* to make it as easy as possible for those who come after you.
I tend to think part of making it easy is not relying overly heavily on fly-by-night cool popular tools (even if the tool is prompted as being a reproducible research tool), and stick to the absolute basics.
I'm guilty of doing this poorly, I probably rely these days a bit too much on tidyverse packages in #rstats when I know perfectly well how to do things with base functions that have been stable since 1998...
I tend to think part of making it easy is not relying overly heavily on fly-by-night cool popular tools (even if the tool is prompted as being a reproducible research tool), and stick to the absolute basics.
I'm guilty of doing this poorly, I probably rely these days a bit too much on tidyverse packages in #rstats when I know perfectly well how to do things with base functions that have been stable since 1998...