[Talk-us] Complex intersection mapping

Martijn van Exel martijnv at telenav.com
Mon Oct 21 23:40:11 UTC 2013

On Tue, Oct 15, 2013 at 10:52 PM, Russ Nelson <nelson at crynwr.com> wrote:
> It *is* incorrect to map in contravention to the description in the
> Wiki.
> It *is* incorrect to map without the wiki containing an explanation of
> how you map.
> If there are a dozen different ways to enter a stop sign into OSM, and
> they're all documented in the Wiki, that's good.
> If there are a dozen different ways to enter a stop sign into OSM, and
> one isn't documented in the Wiki, that's bad.
> If, of course, two people read the Wiki and map differently, that's
> not their fault -- it's the Wiki's description's fault.

Unfortunately, the wiki is not in a state that allows for such rigid
statements connecting mapping practices and documentation. I do agree
generally that everyone should look to the wiki for guidance on how to
map things, but it can't always be the definitive voice you seem to
make it out to be. Let me just give the one seminal example of
confusing, sometimes contradictory information on highway
classification, as represented on no less than four pages here
here https://wiki.openstreetmap.org/wiki/United_States_Roadway_Classification_Guidelines,
here https://wiki.openstreetmap.org/wiki/United_States_roads_tagging
AND here https://wiki.openstreetmap.org/wiki/Highway:International_equivalence.
Another source of confusion on the wiki is the voting system on tags,
which evokes a sense of authoritativeness on decisions made per this
process that does not exist in reality, because so few people actually
take part in this process, and many actively dismiss it as being
misleading and confusing.

Back on topic, I see general support for straightening out
intersections where a road has continuous dual carriageways on both
sides, but Minh's specific cases make a lot of sense to me: we should
not overcomplicate situations and make them less legible and less true
to ground truth:

where I do prefer the first solution.

One followup question I do have is about one of the other examples of
elaborate intersections Minh raised, the Continuous Flow or or XDL
intersections (example
http://www.openstreetmap.org/browse/relation/1284976), I would prefer
to put a no-U-turn restriction on
http://www.openstreetmap.org/browse/node/1002992385 - agreed?

Thanks again for all your feedback.

Martijn van Exel
OSM data specialist

More information about the Talk-us mailing list