When does crossing:markings=surface apply?
Only when surface material is changing or also when only the surface color is changing?
The question come up because of this wiki change
Key:crossing:markings: Difference between revisions - OpenStreetMap Wiki
This is an overly strict and literal interpretation without considering the context. =surface was invented to avoid crossing:markings=solid , which is only color paint.
As proposed, surface was about the surface=* material changing. If the whole crossing is painted a solid color, thatâs what solid was originally proposed for, before we removed it from the proposal as a minor edge case.
The exact text was " only by a change in surface" , rather than surface= precisely. surface:colour= was mentioned afterwards. Color-only should be compatible? Proposal:Crossing:markings - OpenStreetMap Wiki
Another issue nowadays is the âpaintâ can have patterns applied on it, making it resemble a different surface= while itâs not. What to do with those?
But the question that should be asked first is whether the red surfacing has a legal meaning. If it doesnât, it wonât be =solid either, or =dashes;solid;pictogram here.
Slightly off-topic.
Based on the =surface wiki description, shouldnât this crossing be crossing:markings=lines;surface?
Back on topic, I always thought the main difference was colour, not surface, by a quick look of the images. Maybe the image could depict paving stones of a similar colour to make it more clear? Something like:
crossing:markings=surface was basically a compromise: some jurisdictions consider crossing markings to have legal significance but donât consider, say, a low-visibility brick treatment to be a marking. Yet we knew mappers would gravitate toward yes or solid because of the color change, so surface is tantamount to saying, âNo, there isnât really a marking, but see surface=*â. In this case, there are markings â the transverse lines â so lines;surface is a more detailed way of saying yes;no. 
On the other hand, solid started out as a misunderstanding. The whole tagging scheme was originally based on this chart:
This chart was extracted from a 2002 U.S. FHWA guide, which alludes to a study finalized in 2005 that surveyed 1,000 marked crosswalks in 28 cities and 2 counties across the country. Each of the patterns is represented in the study. They counted at least one solid crosswalk but counted brick crosswalks as unmarked. The studyâs authors at the Pedestrian and Bicycle Information Center maintain a non-commercial-licensed site that solicits photos of crosswalks and presumably contains some of the photos from their study, but the closest thing to a solid crosswalk in the U.S. that I could find was this decorated crosswalk.
The pattern is extremely rare, but we originally included it under the assumption that it was roughly as likely as the other patterns in the chart.
Yes, that would be better. I think @1998alexkane was following the artistic style in (slightly older) FHWA manuals, in which solid red wouldâve represented brick paving. The current MUTCD standard has improved upon this design:
Why should a change in surface material be more important than a change of color?
The tag surface:color=* does not say something about if the color is different to rest of the road.
If crossing:marking=surface is only refering to material then I think, we need something different for solid_color.
Donât know, thatâs what is wrote on the Wiki. I always thought colour was the discerning element before this thread was created. But if that is the rule, the image should be clearer.
I think we should keep things simple, most people isnât going to read FHWA guides, studies and MUTCD standars while choosing an icon on iD editor or an image on SCEE.
Well, if crossing:markings=solid is the originally intended way to express the situation, then Iâd say, we should add it to the wiki, and not just as a footnote of crossing:markings=surface.
Love your proposal for a picture of change of surface, by the way. People tend to look at pictures, and only read half-assedly 
Just to make myself sure, how should this crossing (just red paint) tagged then? Or this section here.
The second one is not really crossing, itâs a âAusfahrt ĂŒber einen abgesenkten Bordsteinâ (driveway via a lowered curb). I would not mark this as crossing at all.
Fair, and the first one?
BTW: You can click 10 meters forward, then there are two more crossing for the driveway of the gas station, here and here. The very last one does have dashes, the one before not.
looks also only like a driveway (because there is no give_way sign) but without curb. I would guess crossing pedestrians have also priority because the cars are crossing the sidepath. Then it would be the same like the second one. But Iâm not sure.
Thank you for letting me understand! That means that if there is only a driveway that crosses a shared cycle and footpath, no crossing Node is needed at all. So not here either?
(Mapillary is from the other side only)
There a residential street meets a primary road and the foot-and-cycle path crosses the residential street. If I remove the crossing here, Iâm sure StreetComplete asks how people cross here. Do I then answer âinformalâ?
I think this is a street. You have a stop_sign, a street name but also not the kerb as indicator.
But Iâm not sure in detail with traffic rules in austria.
No it is not crossing=informal - because it is not a crossing and at unmarked and informal crossing the carriageway has priority.
it would be crossing:continuous=yes - the cars have to wait.
Two examples from Germany:
But without any crosssing:markings=* â do I understand that correctly?
in both cases âFahrradfurtenâ along a road with missing dashed lines, the cyclists have priority, the pedestrians not.
For me this was until now crossing=uncontrolled with crossing:markings=surface.
maybe crossing:markings does not hurt, but not with highway=crossing or crossing=uncontrolled/marked.
According to the wiki, crossing:markings requires highway=crossing. I think we should at least add âor crossing:continuous=yesâ, because crossing:continuous=yes is a crossing, just not the footway crossing the carriageway, but the other way around. And sometimes, these footways or cycleways are marked so that even the most unattentive driver gets it.
Agreed, solid was only a footnote because we couldnât find enough real-world examples. As I recall, there was just the one blue crossing on the studyâs website (which we couldnât locate) and a solid yellow raised crossing in an IKEA parking lot. The tagging scheme already allows for Any Tag You Like for one-off exceptions, but if this is an actual phenomenon somewhere, then it deserves a properly documented tag.
Unfortunately, none of the Mapillary links are working for me so I donât have the visual context, but I strongly disagree that a sidewalk-driveway intersection is not a crossing, especially in the case of business âdrivewaysâ and driveways for residential complexes of 5+ units. (Thatâs not a random number, itâs the threshold for (some?) American DOTs)
There should be a tag to differentiate them from the crossings across the primary roadbed (and I really still donât completely understand why crossing:continuous=yes isnât that tag, but I remember getting pushback in the thread for itâŠ) because if I have to worry about my path of travel intersecting with a path of travel of a vehicle, thatâs a duck crossing.
