[Talk-us] Mass Change of Highway Classification in Larimer County Colorado

[Brian M. Sperlongano]

The whole point of the highway classification values is to capture the
hierarchical network of roads.  Renderers rely on the hierarchical nature
of the classification values to make decisions about which zoom levels to
show different roads at.  Trunk/primary/secondary/etc, are not, repeat, ARE
NOT, proxy tags for the number of lanes on a road, whether or not the road
is divided, the speed limit, or the presence of stop lights, shoulders,
intersections, or sidewalks.  There are separate tags for all of these
things.  Don't abuse highway classification for tagging the fact that a
road has high speeds or lots of lanes.

Regarding:
> roughly subjective based on network, size and character like we have
been.  It works.

No, it does not work, and this attitude is why highway classification
tagging in the US is (a) cartographically broken, (b) has no defined
standards that mappers can use and (c) results in constant edit wars and
hurt feelings (the origin of this thread being case in point).

Let me show you why it's broken.  Take a look at the OSM map at zoom 7
around Oklahoma[1].  Mappers have changed the highway classification of the
roads in that area based on their physical characteristics.  At zoom 7, OSM
shows trunk and above, and thus US 50 and US 54 pop into and out of
existence based on the whims of local mappers.  The result is an entirely
avoidable cartographic dumpster fire.

In comparison, our competitors[2,3,4] keep routes together, which results
in route continuity at all zoom levels, *regardless of physical
characteristics*.

[1] https://www.openstreetmap.org/#map=7/36.829/-101.285
[2] https://www.google.com/maps/@36.7823325,-99.7599754,7z
[3] http://mapq.st/2dypPbv
[4] https://www.bing.com/maps?cp=35.95416~-99.784913&lvl=7

In order to solve this, we need to let go of the idea that physical
characteristics have any bearing on highway classification.  I don't care
that there are examples of state routes that have higher speed limits and
more lanes than US highways in some cases.  I don't care what NE2 tagged in
the past.  At a map zoom level that shows multiple cities, it should show
the major routes between the cities, regardless of the quality of those
roads.  At higher zooms, the map should show the major routes between
smaller towns and cities, and so forth.

For this reason, US highways should always be trunk, regardless of their
physical characteristics.  The reason they are US highways is because they
provide long-haul connectivity between cities and they should appear at
lower levels of zoom than other roads, based on their importance in
connecting cities.  Likeways, state routes should be at the next level of
importance, followed by county roads in places where those exist.
Subjective judgements about road network connectivity importance should be
limited to the lowest classification categories where we don't have routes
to go by.

I would advocate for US definitions that look something like this:

motorway:
- Signed interstate highways
- Any other controlled access, grade-separated, divided highway built to
interstate or near-interstate standards for a significant length

trunk:
- US highways

primary:
- Numbered state highways

secondary:
- For states that have them, county roads; otherwise, the most important
named roads that provide connectivity between higher classification roads
and lower-classification roads.

tertiary, unclassified:
- Lesser named roads that provide connectivity between secondary roads and
residential/service/track roads, where the more important roads are
tertiary and the less important ones are unclassified
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-us/attachments/20210504/75341cb0/attachment.htm>