Connected ways rendered as Unconnected ways due to Garmin low resolution
Hi All, Minor, but irritating and hard to find issue. I have searched in Archive for "Unconnected ways", but I cannot find any relevant postings on the subject. Issue: in JOSM we can and should map accurately. A roundabout is a circle (Create Circle), a building is a rectangle (Orthogonalize Shape). Looking at the Map in https://www.openstreetmap.org/ it's perfect, in OsmAnd it's rather oké, but MkgMap on Garmin Device (or BaseCamp) is a "sloppy" polygon. More important: connected highways are sometimes (sometimes!) rendered as not connected. Therefore routing is not possible because of "Unconnected ways". These exceptions are hard to find (for me it is). Validation Tools (JOSM, Osmose, OSMI, Keep Right) won't find these errors because on OSM it's oké. One example: <http://gis.19327.n8.nabble.com/file/t344065/MKGMAP_NOT_Connected_highways.png> To resolve (IF I do find an exception): - and/or: Simplify ways, most specific for a roundabout (8 nodes is preferred to 16 nodes). - and/or: Move the Connecting node (which is Unconnected on the Map) very slightly. Small, minor move is always enough to resolve. Also: Rendering can resolve. If the issue does occur on a bicycle map (highways are rendered by narrow lines) than the issue will not occur if the width of a highway is rendered larger (on a Motorcar/general map). I'm prepared for an answer like: "This is Garmin, not MkgMap!" Or is something like "snap to grid" an option for MkgMap to resolve by a work-around? Or improved logging of exceptions? If I can find these "errors" I can resolve! Most annoying (for me): my attempts to improve card quality leads to "routing errors". / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html
Hi Eric, see http://gis.19327.n8.nabble.com/Distorted-lines-tp5831842.html and esp. my "bed of nails" explanation: http://gis.19327.n8.nabble.com/Distorted-lines-tc5831842.html#a5831902 maybe you can give us an OSM link to the crossing shown in your example? For me it looks more like an error in OSM. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von AnkEric <eric_internet@casema.nl> Gesendet: Mittwoch, 29. Januar 2020 11:28 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: [mkgmap-dev] Connected ways rendered as Unconnected ways due to Garmin low resolution Hi All, Minor, but irritating and hard to find issue. I have searched in Archive for "Unconnected ways", but I cannot find any relevant postings on the subject. Issue: in JOSM we can and should map accurately. A roundabout is a circle (Create Circle), a building is a rectangle (Orthogonalize Shape). Looking at the Map in https://www.openstreetmap.org/ it's perfect, in OsmAnd it's rather oké, but MkgMap on Garmin Device (or BaseCamp) is a "sloppy" polygon. More important: connected highways are sometimes (sometimes!) rendered as not connected. Therefore routing is not possible because of "Unconnected ways". These exceptions are hard to find (for me it is). Validation Tools (JOSM, Osmose, OSMI, Keep Right) won't find these errors because on OSM it's oké. One example: <http://gis.19327.n8.nabble.com/file/t344065/MKGMAP_NOT_Connected_highways.png> To resolve (IF I do find an exception): - and/or: Simplify ways, most specific for a roundabout (8 nodes is preferred to 16 nodes). - and/or: Move the Connecting node (which is Unconnected on the Map) very slightly. Small, minor move is always enough to resolve. Also: Rendering can resolve. If the issue does occur on a bicycle map (highways are rendered by narrow lines) than the issue will not occur if the width of a highway is rendered larger (on a Motorcar/general map). I'm prepared for an answer like: "This is Garmin, not MkgMap!" Or is something like "snap to grid" an option for MkgMap to resolve by a work-around? Or improved logging of exceptions? If I can find these "errors" I can resolve! Most annoying (for me): my attempts to improve card quality leads to "routing errors". / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Hi Gerd, I was very afraid you would ask for an example! I'm 100% sure it's OK on OSM. JOSM validation would protest! Immediately! Examples are very rare AND are always Resolved. If I find one piece of my own Garbage tomorrow I will resolve the same day. The next day I will conclude that it has indeed been solved. By exception: I will wait for OpenFietsMap to see if it has the same issue. In case I do wait, the issue is always confirmed by OpenFietsMap. Also: if I resolve, I will not (no need) Connect or JOIN. My only actions: move the connecting node (and sometimes: delete some nodes). I could have a look at my Changeset Comments. And than... revert my solution? Create Errors deliberately? Lol, please don't tell anyone; -)) Back in memory lane: I now do remember one issue. But also: that was one year ago! The Netherlands, Wassenaar? Roundabout? Can I search my own Changeset Comments? But also: you do not ignore my issue. So next time I do find one, I will NOT resolve and send you the link immediately. Next: I will read your "Distorted-lines". But be prepared for ignore modus: I might not respond anymore because I cannot find an example. Now I understand why it took a few years before I posted this. If it is not reproducible, a solution is almost impossible. / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html
Hi Eric, it is rather easy to reproduce for me because I can add some debugging code which shows the effect of the rounding to garmin units. You can try yourself: Draw some buildings in JOSM and create copies at other positions. The biggest distortions happen when points are near to the center of 4 nails in my "bed of nails" explanation. I asked for an example because JOSM will not report such a situation. It only complains when a road ends near another one. Try attached sample. Gerd ________________________________________ Von: mkgmap-dev <mkgmap-dev-bounces@lists.mkgmap.org.uk> im Auftrag von AnkEric <eric_internet@casema.nl> Gesendet: Mittwoch, 29. Januar 2020 12:06 An: mkgmap-dev@lists.mkgmap.org.uk Betreff: Re: [mkgmap-dev] Connected ways rendered as Unconnected ways due to Garmin low resolution Hi Gerd, I was very afraid you would ask for an example! I'm 100% sure it's OK on OSM. JOSM validation would protest! Immediately! Examples are very rare AND are always Resolved. If I find one piece of my own Garbage tomorrow I will resolve the same day. The next day I will conclude that it has indeed been solved. By exception: I will wait for OpenFietsMap to see if it has the same issue. In case I do wait, the issue is always confirmed by OpenFietsMap. Also: if I resolve, I will not (no need) Connect or JOIN. My only actions: move the connecting node (and sometimes: delete some nodes). I could have a look at my Changeset Comments. And than... revert my solution? Create Errors deliberately? Lol, please don't tell anyone; -)) Back in memory lane: I now do remember one issue. But also: that was one year ago! The Netherlands, Wassenaar? Roundabout? Can I search my own Changeset Comments? But also: you do not ignore my issue. So next time I do find one, I will NOT resolve and send you the link immediately. Next: I will read your "Distorted-lines". But be prepared for ignore modus: I might not respond anymore because I cannot find an example. Now I understand why it took a few years before I posted this. If it is not reproducible, a solution is almost impossible. / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
I did find one 3 year old example: <http://gis.19327.n8.nabble.com/file/t344065/Changeset-46044991.jpg> https://www.openstreetmap.org/changeset/46044991 (AnkEric, 2/3/17) Changeset Comment: {highway=tertiary_link} (id:338819633) did NOT connect to Kanaalkade by mkgmap Renderers for no obvious reason. Trying to resolve by mapping DIFFERENTLY, not better or more correct. By (temporarily) reverting my Changeset-46044991 you should be able to verify: 1. OSM was OK before Changeset-46044991, all ways are connected. No Errors, no warnings; 2. BUT: Rendering was not OK (see: image OpenFietsMap); 3. In my Changeset-46044991 I did not FIX or CORRECT errors. I only moved some nodes; 4. Resolved by Changeset-46044991 (you just have to believe me on this); 5. In my Today's Map it's Rendered OK ("proof" 3 years after the fact). <http://gis.19327.n8.nabble.com/file/t344065/Rendering_Changeset-46044991_OK_date_29-01-2020.jpg> Better example would be a highway (not) connecting to a roundabout. But I cannot find an example. Possible Mapping Errors in OSM: Crossing highways/Way end node near other highway/Highway duplicated nodes. JOSM validation will warm for all 3 possible errors/warnings. <http://gis.19327.n8.nabble.com/file/t344065/JOSM_validation_1x_Error_2x_Warning.jpg> In my experience: OSM is OK, no Errors or Warnings, but the next day (Overnight Map Making procedure) the Connected ways are Rendered AS Unconnected. I fact the connecting Node (1 node!) is Rendered twice on different map locations: 1x as member of roundabout, 1x as member of the (un)connected highway. Oohhh, by the way, I don't understand any of this; -) / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html
Today I found another example of Unconnected ways. But to simplify the issue: - On my BIKE (bicycle) MAP the ways are Unconnected. - On my HIKE (foot) MAP the ways are Connected. Same OSM Data, same date (February 19, 2020), same mkgmap-r4452. I think, believe, I do see: Unclassified (south to north) is identical for both maps. But the secondary (west to east, gray) is relocated (bend) to connect to the unclassified in Hike Map. All crossing ways are connected correctly. No recent updates, since I wanted this example to be left unchanged. <http://gis.19327.n8.nabble.com/file/t344065/Bike_Unconnected_Hike_Connected.png> The scripts bicycle vs. foot are almost identical, same script except for some: If - Then - Else Statements selecting either "access_bike" or "access_foot" as include. if (map_type=1) then (highway=cycleway) & (sidewalk ~ '(both|left|right|yes)') {set highway=footway} (highway=cycleway) & (segregated=yes) {set highway=footway} end And on Hike Map I do render street-names, on Bike I don't. If (... IF...!) you want to look into this - again - I could run both scripts in debug mode: highway=* & osmid()=6016xxxx { echotags "BEGIN" } highway=* & osmid()=6016xxxx { echotags "END" } Attached OSM: Bike_Unconnected_Hike_Connected.osm <http://gis.19327.n8.nabble.com/file/t344065/Bike_Unconnected_Hike_Connected.osm> BTW, FWIW: OpenFietsMap (15-02-2020) is oké. I have seen this before: sometimes Me and OFM share the same issue, sometimes not. / Eric (AnkEric) -- Sent from: http://gis.19327.n8.nabble.com/Mkgmap-Development-f5324443.html
participants (2)
-
AnkEric -
Gerd Petermann