[Tagging] Describe explicitly that values of highway tag do not imply anything about road quality (except highway=motorway and highway=motorway_link)

[Ilpo Järvinen]

On Wed, 19 Aug 2015, Mateusz Konieczny wrote:

> I am planning on adding this information to wiki in near future.
> 
> Note that it afffects also highway=path - in that case some/many
> people are convinced that there is some kind of difference of implied
> quality between highway=path and highway=footway.

Funny, how you put it as informal paths and unpaved, non-made forest 
trails so far have been pretty much what I expect throughout the world 
(although I admit I've no experience on areas in Africa). But surely you 
can pretend that forces of erosion and weather, and the effects of a 
similar terrain and vegetation suddenly work differently somewhere :-).

But of course once shiny built-up footway can fall into a state of such 
disrepair that is in worse condition than those informal paths, so you 
have some truth in what you say too.

> highway=path wiki page already contains information that
> highway=path "may have any type of surface"

Most of those who would want to tag those "paths" would likely prefer 
introducing own highway type anyway (such as highway=trail as =path is 
unfortunately taken) - but in that case there's always fierce opposion 
from few who are convinced such "paths" don't exists. And that 
non-existance of alternative to highway=path is likely the main reason
why they have been forced to use highway=path in the first place (even 
though they know some people think that =path == =footway).


--
 i.


> 2015-08-12 10:56 GMT+02:00 Mateusz Konieczny <matkoniecz at gmail.com>:
>       What is below is supposed to document existing consensus rather
>       than
>       introduce new rule. I am aware that there is significant
>       majority
>       convinced that some road types universally guarantee or imply
>       certain
>       values (like widespread usage of highway=track to mark all
>       unpaved
>       roads and only unpaved roads, popular in some regions).
> 
>       I am also aware that in some regions "dual carriageways are
>       usually at
>       least highway=tertiary", "paved roads are at least
>       highway=unclassified", "highway=primary has at least 2 lanes in
>       each
>       direction" or "paved roads passable for the entire year are
>       highway=trunk" are good approximations - but it does not make it
>       universally true or necessary requirement to use these tags.
> 
>       proposed description to be added on
>       http://wiki.openstreetmap.org/wiki/Key:highway - maybe also on
>       pages
>       about individual road types.
> 
>       Only highway=motorway/motorway_link implies anything about
>       quality.
>       Other road types, from highway=trunk through highway=tertiary to
>       highway=residential=residential/service or
>       highway=path/footway/cycleway/track do not imply anything about
>       road
>       quality. In area with poor infrastructure road forming main road
>       network, of the highest importance in region should be tagged
>       highway=trunk - no matter whatever is is high-quality asphalt
>       road or
>       unsurfaced tract unusable after major rains.
> 
>       In absence of surface, tracktype and other tags describing
>       quality of
>       road one may try to extrapolate this information from
>       value of highway tag. Note that this needs tuning for every
>       region.
>       Typical highway=primary may be drastically different in various
>       places
>       across the planet.
> 
>       It is highly recommended to add and use tags describing road
>       quality
>       like surface. In one region it may be obvious that all
>       highway=tertiary
>       are paved and highway=track unpaved but no assumption like that
>       will
>       work worldwide.