I agree with you, differences in pronunciation, cadence, etc. should be taken in account as well. Though measuring those could take longer, if possible.
Spanish native here, confirming that RSA is the institution that sets the language standard.
But, people always deviate from it, though in my experience in word meanings and pronunciation, never in grammar to a degree that it become intelligible to another Spanish speaker.
The toughest film to listen to for me was "The rose seller"[1] (1998), took me like 10m to get my ear accustomed to their pronunciation.
What always did, going through awesomewm, stumpwm and xmonad to now sway, was to used whatever of those services were available from the OS. Some I had to launch by hand, but that was a one time change always (some of them across WMs). Nowadays it is even easier, since most of what I needed to launch manually are now user-level service from systemd.
In the end, I am running most of the normal services a GNOME desktop would be running, but without using the GNOME shell, and that has worked very well.
[1] https://github.com/mickeynp/combobulate