[Tagging] RFC: pedestrian crossing as an area

[stevea]

On Jan 13, 2022, at 9:01 AM, Brian M. Sperlongano <zelonewolf at gmail.com> wrote:
> Until there is some reasonable critical mass in terms of data consumer support and feature proliferation in the map, I think we should avoid approving tagging schemes for nano-mapping, and instead allow such mapping to continue growing organically in order to learn what innovations result from it.  The use of nano-mapped features is still in its infancy, and I suspect it would be a mistake to approve tagging for it until we're able to fully embrace how these features fit into the big picture and how the data consumer needs stack up against them.

+1.  I’ve called these sorts of tagging “micro-mapping” (leaving nano- for very tiny features, like individual drinking fountains, vending machines…), but we’re talking the same language.  I see “crosswalk areas" as much like public_transport, where maybe in one country with particular pavement markings and/or signage, a certain flavor of tagging emerge, then it crosses borders, becomes highly complex as it varies, and OSM needs a fairly rich "sub-language” to describe the semantics.  This latter process is what caused PTv2 to necessarily emerge from a simple (underspecified, but working, albeit limited) PTv1.  PTv1-> PTv2 was not easy, has taken many years and as it continues to deploy, will continue to take many years to be fully established.

OSM can (should) learn from this and exactly as Brian suggests, allow this sort of real-world item to develop its own complexity.  Such a thing will vary from country to country and need to be well-modeled with a deeply-specified tagging scheme if OSM has any hope of properly capturing all the various flavors.

And while, to me, this isn’t a burning issue, I’m cautious at speculating on the “usefulness” of any set of real-world objects OSM might map:  some could find this laughably trivial and ridicule its syntax development, while others (perhaps disabled pedestrians?) could find this an absolutely essential use-case.

As we have seen, the global perspectives necessary for multiple-country differences do not always make this easy, though I believe OSM is up to the task.  The solution-space size of this being relatively small (compared to, for example, PTv1 and PTv2 differences), it could even be a good pedagogical case for how OSM might develop such tagging schemes going forward.  Maybe “develop in one country, expand to another, look towards global deployment…” as a start.  I wouldn’t want the brittle, sort of broken “re-invent the wheel” (extant in PTv1->PTv2) experience to be repeated any more than necessary, if at all.