-
Notifications
You must be signed in to change notification settings - Fork 823
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Clean-up paths and tracks, in particular on z13/z14 #747
Conversation
Note that on z13/z14, the paths are very faint, but that's not a problem. Short paths are not necessary to see at that zoom level, and long paths will be visible even if they are faint. |
- Make paths and tracks less visible on z13/z14, by making them narrower and remove the casing at these zoomlevels. This declutters in high-density environments, while at the same keeping the paths and tracks visible in low-density environments. - Hide private paths and tracks on z13/z14. - Define path/track widths with variables. - Make widths of path/track bridges and tunnels more consistent. - Add rendering for steps in tunnels. - Add background (glow) to steps, just like footways. This solves the following issues: * gravitystorm#211 (Poor result of rendering areas with many footways in zoom level 13) * gravitystorm#620 (Tracks too dominant on z13) * Trac 1508: don't render tracks of high tracktype at low zoom (-> wontfix) * Trac 3788: Don't render highway=track, access=private at z13&14 * Trac 4015: Rendering of highway=path with access=no is inconsistent * gravitystorm#634: private footways should not be rendered up to z13
See also #668 that proposed less elegant way of solving some of mentioned problems by not rendering footways at z13. |
Are you planning to provide more before/after? I am especially interested in whatever Kraków at z13 will become useable after these changes. |
I think in general it should be ok even without more before and after. I think we can get the idea. |
Nice work @math1985 |
Indeed it is. |
Thank you, it looks great :) |
I suspect in most parts of the world (i.e. away from densely populated areas) this will have a significant negative effect. Here's an example in Derbyshire at Z14: The width of that image is about half an hour's walk - definitely the sort of scale at which it should be possible to see how to get from A to B. |
Overall this looks much better. The footways are still visible. I am unclear what you mean about not being able to see the routes you want to walk in that area. |
To be precise - when looking at the above area on a mobile phone screen at Z14, I can't see the footpath that goes from the wood west of Kelstedge to the road to the north at all (even indoors out of the sunlight). On the picture above I can see some hints of pink (because it's a much bigger screen, and because I know it's there to look for), but I couldn't use this map as an actual map for navigating with. |
Well mobiles have a tiny screen and it is an issue in general. As writing from my tiny smart phone you just have to zoom in sometimes. |
I see what you mean, @SomeoneElseOSM . Footways that go over farmland are even harder to see at z=14, see https://www.openstreetmap.org/#map=14/52.9232/-0.8588 I don't know if it still could be tweaked a bit. Maybe a darker colour (but still no "halo")? There has to be a compromise made for the default map. Before, the footways were much to important at that zoom and really made the map hard to read in dense places. Now, you have to zoom in one more step if you can't make out the footway. That's OK for me. For hikers, there are special styles available, even the cycling mapstyle shows the footways much stronger. |
@SomeoneElseOSM On my screen (laptop) footways are visible. And in areas with many footways map is no longer a reddish blob of footways, so IMHO it is a success. |
Re the screen size comments above, I'd suggest that designing for traditional computer screens is very much a first-world-centric view (see for example http://www.pewglobal.org/2014/02/13/emerging-nations-embrace-internet-mobile-technology/ ), and is becoming the default internet access option in developed countries too. One of the big plusses of the "big OSM site redesign" last year was how well most things now work on a mobile phone browser screen, but this change has been a major regression in that area. The "reddish blob of footways" (such as the Polish graveyard) is very much an edge case. Most of the world isn't like that. Most of the inhabited world DOES consist of footpaths and tracks running through forests, fields and suchlike. Surely OSM should be for everyone, not just those rich or lucky enough to be sat behind large desktop or laptop screens? |
@SomeoneElseOSM |
@SomeoneElseOSM Thank you for your feedback. This is the area you are referring to (when referring to a particular area, please provide a link as well next time, makes testing easier). I tested that area on my phone, but on my phone (Ace 3), I didn't have problems seeing the footway, although I admit it doesn't particularly stand out. At the moment, we have no way to let the rendering depend on the density of the surrounding features. In a city, footways are less important on z14 than in the countryside. Therefore, we need to use a tagging that works both in the city and in the countryside - an explicit design goal of this change. As you know, the previous rendering did not work in cities. The current rendering is the best compromise I could find. If you or anyone else has suggestions to tweak the rendering to make it more suitable for the countryside, feel free to submit a pull request. |
I'd disagree that "the previous rendering did not work in cities". A couple of edge cases (like the Polish graveyard) have been highlighted, but it was not a general problem. Here's an example of a mostly fully-mapped city, under something very like the previous rendering (some footways are rendered as paths): and here it is now: (http://www.openstreetmap.org/#map=14/53.2241/-1.4417 for info) Previously, you could see that a path ran north from Selby Close; now, not so much. What we have currently isn't in any way a compromise. It makes the map look prettier, but less useful as a map. |
This was certainly a general problem, and not restricted to Polish cemeteries. See for example the before-and-after here. |
@math1985 Could you explain what was the problem with your map of Breda that you sought to fix? All I see is information (that there are footways linking residential streets) removed from the map. (current rendering http://www.openstreetmap.org/#map=13/51.5861/4.7674 ) |
In my opinion, the footways were rendered much too prominent for what they really are. They also hid the colour of the background. |
Sorry - I'm confused - exactly what information was hidden by the more prominent display of footways on the previous map of Breda? What can I use that map of Breda for now that I could not before? I've lost the ability to route on foot from one side of town to the other; what have I gained? |
No, you never had that ability in the first place. You could only see some red blur, but had no idea about which direction the footways went.
The map is easier readable for car drivers, and it is eaiser to distinguish parks and cemeteries from residential areas. |
Can you give me a specific example? I can't see a single park or cemetery that was easily confused with a residential area area before, and more recognisable for what it is afterwards. I'm not being obtuse here; just genuinely trying to understand the reasoning behind the change. As for car drivers; isn't Mapquest Open more targeted towards their needs? http://www.openstreetmap.org/#map=13/51.5850/4.7569&layers=Q |
Berlin is a better example: http://bl.ocks.org/tyrasd/raw/6164696/#13.00/52.5279/13.3808 Allotments, parks, some residential areas are a red blur. In Kraków problem was not limited to cemeteries - allotments, parks and residential areas also were heavily affected: http://bl.ocks.org/tyrasd/raw/6164696/#13.00/50.0753/19.9773 At least for me the main problem was that map was extremely ugly. |
All the other road types scale down in prominence of rendering as zoom decreases, but footway wasn't. |
@mkoniecz I'm afraid I'm simply not seeing the improvement. Taking the Tiergarten in Berlin as an example, previously I could see that there were paths that ran north and south of the river, and north and south of the main road through it. Now http://www.openstreetmap.org/#map=14/52.5155/13.3618 it isn't clear that that it's publically accessible at all. There was plenty of green visible in the old Tiergarten rendering, so it was very obviously a park of some sort, but in addition I could say that it was a place where I could go for a walk. Now that information has been removed, with no extra information conveyed.
This I suspect is the fundamental difference of opinion here - you'd like a map to look nice; I'd like a map to use as a map. Both are valid aims I guess, but when people visit osm.org which are they likely to want to see - a piece of art, or a useful map? |
Someoneelse, All these publicly accessible paths are still viable they are just thinner. The concept of zooming in well established for users of any map to see more detail of features. |
IMO we should keep z13 as is now. |
"Taking the Tiergarten in Berlin as an example, previously I could see that there were paths that ran north and south of the river, and north and south of the main road through it." It still visible, and fortunately group of footways in park is no longer more visible than nearby highway=secondary. And yes, it may be a good idea to make footways and cycleways at z14 more prominent. |
Some ideas:
|
I think at zoom level 14 and 15, it would be best to render all footpaths the same, and not have a special render for footpaths with any bridge tag. |
remove the casing at these zoomlevels. This declutters in high-density
environments, while at the same keeping the paths and tracks visible in
low-density environments.
This solves the following issues: