which metrics?

[mpadge]

here you go @mdsumner - this should be very quick to assemble. We can just start with vectorized C++ implementations of:

1. Haversine
2. Vincenty
3. Mapbox cheap, as explained in this blog post.

This will all be able to be done quickly with only Rcpp dep, but we could even dump that and then be essentially dependency free?

[mpadge]

And by the way: available::available("geodist") gives the all-clear

[mdsumner]

I'm happy with these and Rcpp, for now Rcpp is basically an assumed basic most of the time!

[mdsumner]

I'm inclined to default to metres as the basic assumed unit as it is with projections - do you feel strongly for km?

geosphere is metres but others km (or miles :)

x0 <- cbind(0, 0)
x1 <- cbind(0, 1)
## m  
geosphere::distGeo(x0, x1)
# [1] 110574.4

## km 
spDistsN1(x0, x1, longlat = T)
# [1] 110.5731

## km 
fields::rdist.earth(x0, x1, miles = FALSE)
# [1,] 111.3239

I think we (royal) tend to work in m/s as a default rather than km/h - is there a stronger established convention outside of R?

EDIT: it seems cheap ruler uses km default so maybe I need to suck it up

[mpadge]

I'm easy on that, but one advantage of default km would be an ability to offer miles as well. That'd require another param, so we may as well just add units and enable ask options, right? I can't really see us facing much of a parameter overload issue

[mpadge]

closing this now that the big three have been implemented, and opening a units issue to finish the job