I feel like the tagging for surfaces is insufficient if the split is parallel i.e. along the way instead of lateral i.e. crossing the way.
The general consensus is to simply tag the worse surface but this IMO doesn’t do much justice for cyclists and wheelchair users because sometimes, the worse surface implies it’s a bad way to travel even though it may very much be actually the more preferable option over alternative routes.
It’s consistently paved with paving stones on one side and asphalt on the other but is not a separate foot- and cycleway but a shared one. Even then, when I cycle on it, I almost never driver over the paved part, only on the asphalt one.
For smoothness, the consensus is to tag according to the best part of the way, see Key:smoothness - OpenStreetMap Wiki However for surface, there’s no ranking of what is a better or a worse surface.
The left side of this way is surface=sett, by the way.
For that matter, the reason I made a thread in General Talk and not in Help and Support is because the surface tagging with a split along the way is pourly defined (in the Wiki) and wanted to ask there first (this is on top of the fact that I’m new to proposals and I wanted to sort a few things out first) as well as create a poll on that matter after enough discussion (or someone else, feel free to do so yourself).
Furthermore, thinking about it, the “correct” tagging for split surface ways gets even more complex when you include roads for vehicles as well as ways for pedestrians and cyclists, considering the latter typically don’t feature any lanes at all outside of separated foot- and cycleways (and the tags for them already exist). This is why I feel like @Kovoschiz’s approach is fitting in both ways (pun unintended) albeit for different types of highways.
In the above example, I would use *:left and *:right as the way is a shared path and don’t consider the separation to be lanes either.
However, for roads which serve vehicles like private cars and trams, a separate surface can be considered a separate lane depending on the width of the surfaces (which is common with exclusive but non-kerb-separated tram lanes, though themselves not very well defined), so I’m leaning towards a lane approach (i.e. surface:lanes=*) for them.
But again, this is just me and not what others are thinking so I first want to collect possible alternatives (however unlikely they are to pass) for split surface highways.
