mcsiple/mmrefpoints

Think about dependencies and where they can be reduced

Closed this issue · 3 comments

This package has a lot of dependencies relative to others, and it took a very long time for me to install because of this. In addition to 30 first level dependencies (e.g. the ones in DESCRIPTION), you have 165 total dependencies once you factor in their dependencies.

To visualize this, I ran:

install_github("falo0/dstr")
library(dstr)
dstr()
plotdstr()

image

Lots of packages!! One quick way to limit this is not requiring the entire tidyverse() as it is very large, but only the packages you use (e.g. dplyr). I also don't think you need to include things like here or golem since they may have been used to build your app or in your day-to-day workflow but aren't necessary for users to run the app.

I've gotten it down to 135 packages... I'm going to try to whittle down many more of them. There are quite a few that are used by shiny to make things "production-grade" (e.g., add tooltips for help) so there may still be a lot even after I've simplified it a lot. I did, however, get rid of the full tidyverse dependency (yay!), got rid of here, and trimmed off some others (I guess I still need golem for something but I'm not sure what). Still working on this one.

Rplot

I've gotten it down to 134 (lol) with commit a1a90fd and I'm calling it! I don't know which of these I can do without... we'll see what reviewers say!