New approach to local routing?

[Robinlovelace]

In #481 I am looking into removing the sfNetwork function and associated R/C++ code. Long term, the plan was to integrate the increasingly mature sfnetworks package into {stplanr}, or rather to use it as a 'back end' for local routing. I'm making this a discussion not an issue because I'm not sure what approach is best and would appreciate feedback. route_local() worked great, but was a bit fiddly at times. I would like to reinstate that function and import/suggest sfnetwork. Thoughts?

Directly asking @richardellison (who wrote sfNetworks functions and showed what's possible, is it worth creating a mini package to highlight that functionality or even to update the code?) I suspect it will be less work and better performance to switch to sfnetworks from small benchmarks but cannot be sure it's always faster.

Heads-up @luukvdmeer, @agila5 and @loreabad6 on the {sfnetworks} side, the package goes from strength-to-strength. Heads-up @temospena, for interest in local routing, and anyone else interested in this discussion please chip in!