[Tagging] place=neighbourhood vs landuse=residential

[stevea]

In Santa Cruz, California (USA) [1], for at least one of our neighborhoods (Prospect Heights), there was an ongoing process of an encouraging OSM tagging evolution.  Quick background:  the City of Santa Cruz in OSM is overlayed (all the puzzle-pieces snap together neatly) with landuse=* polygons (residentaial, commercial / retail, industrial...).  The residential ones are accurate, as to extent — this from someone who has lived in Santa Cruz most of my life — though they are more all-encompassing than they actually should be because they are the easy-to-draw all-encompassing polygon.  They are intended, and work well as a "first draft" of landuse in the City, and when all residential, commercial/retail, industrial, schools and parks are included, it adds up to "the whole City" except for state highway corridors.  These large residential "zones" include things like sidewalks, roads, drainage ditches / culverts, ROWs and not necessarily "only the residential areas," (houses and their front, back and side yards) more like "all of the area enclosed by them."  However, in Santa Cruz' Prospect Heights neighborhood [2] in the twenty-teens (and I recall well into the 2020s, too) OSM found several "exact-sized" landuse=residential polygons sprouting up.  Like cut-outs from a plat map or tax map where "only the houses and their lots are."  Strictly residential and nicely precise with pretty curves, excluding street and sidewalks.  Our local county wiki (which I recently updated to say these have vanished from OSM recently) was on track to suggest that when that large, sloppy-but-OK-as-a-first-draft landuse=residential + name=Prospect Heights polygon was full of these deliciously-precise residential-only polygons, the landuse=residential tag on the big enclosing polygon could then have its tag changed to place=neighbourhood.  I wonder about those now-deleted more-precise residential-exactly polygons...that was an encouraging improvement to our data.

It's complicated out there.  Data come, data go.  Heck, even tagging schemes change as they evolve (rare, but it happens).  DaveF has got it right, though deleting a name tag when it duplicates something else surrounding it seems like a lot of assumptions to make (can that broken bond can always be kept by assumptions with certain logic?), so Dave might need to further convince me on that before I'd start to do it myself.  But I don't see that (particular example) as I map around me, that's a new one on me.  See, many of us come up with odd occurrences, even as they are different ones.  Good dialog here, even great.

[1]  https://wiki.osm.org/wiki/Santa_Cruz_County,_California
[2]  https://www.openstreetmap.org/way/220344508