[Tagging] Use of highway=track vs highway=service cemeteries, parks, allotment gardens, golf courses, and recreation areas

[stevea]

Quoting me:
> Contrast that with the older style that is more strictly "highway=service MEANS paved" and "highway=track MEANS unpaved."

On Feb 25, 2021, at 11:35 PM, Martin Koppenhoefer <dieterdreist at gmail.com> wrote:
> I am aware it was done like this in some areas, but it wasn’t accurate even 13 years ago when I started to map. tracktype=grade1 for paved tracks was already there back then.

Mmmm, I'd say "almost."  While it WAS done, (and still is? — naughty!) it was a bit more difficult back then to know WHAT was truly "accurate" or not.  I've been OSM-mapping (and wiki-reading, wiki-writing) for almost that long and while the wiki was extant (and even very helpful back then), things were a bit "rougher," with consensus not quite as clear as has emerged today.  And because we had tracktype=* but NOT surface=*, I know I used tracktype much more often long ago than I do now, though I do still use it, because it is a useful designator AND it renders with "pleasing semiotics" (my opinion) in Carto.

In my mapping, grade1 was always a "more rare" tag, it was for those unusual (in my experience) "truly UNpaved roads/tracks that are so close to paved they approach usability as if they were paved."  As I've said, they might be quite wide, quite smooth, quite straight, connect two sizable towns, be perfectly travel-able in a passenger car with no need for high clearance or 4WD and allow speeds up to maybe 70 km/hr (~45 MPH) or so.  In my mapping, such surfaces ARE unpaved, but are "virtually paved."  Well, those are the roads where I've applied tracktype=grade1.  I now see in our wiki that what looks like "light sealing" (a form of paving, though thin and mild) is what is depicted on a forest road for tracktype=grade1.  The upshot:  in some cases, especially for mildly-hazily-defined tag / semantic parings (like tracktype=grade1 used to be before that picture was included in the wiki), OSM does experience some mild drift after so many years (10, 12, 15...).  And highway=track with "now better" tracktype=* (but used to be not as well-defined) is certainly one of them.  I believe this has a lot to do with why many of us here and now continue to either "settle" or want to on "perfectly defined" semantics.  And why more longer-term mappers (like us) seem a bit shrill in repeating (um, re-stating) our perspectives.  This sort of "semantic drift" happens.

> The typical physical appearance of tracks really depends on the area, in some regions most longer tracks are paved, in others there are hardly any paved tracks.

True, that.  Where "there are hardly any paved tracks" (itself a mildly newer concept that applies to tracktype=grade1, in my experience as I state above) is a factor, this makes it difficult to have a large enough set with which to distinguish such subtleties.

Wow, this is a rich, deep, subtle topic, that has experienced some "smearing" over the years, making things difficult, it's true.  While it's important to note that there remain in the map older methods of tagging (from older wiki, older methods of understanding, because of fewer examples and lacking in modern tags like surface=*), I think it's more important to emphasize what we agree upon to best tag going forward.

SteveA