Well it depends, right? My most recent Python packaging adventure was managed through Docker, but this was because I also had Java-based dependencies (reAgent R

Well it depends, right? My most recent Python packaging adventure was managed through Docker, but this was because I also had Java-based dependencies (reAgent RL framework). Given how much easier that was than getting people to use conda/pip (conda is better for DS stuff as it handles C-level dependencies), I completely understand how people just suggest Docker. Like, if you're doing pure Python lots of this may not be necessary, but as soon as you start having multiple C-level dependencies pip breaks (and pip didn't even do dependency resolution till last year(!)).