mergeroads sometimes breaks routing

I found a very strange routing behaviour connected to the mergeroads-patch. The Problem area is here: http://www.openstreetmap.org/#map=17/47.89307/10.63390 When I try to route ( Automobile-mode, minimize-time ) from corner Liegnitzer-Straße/Glogauer-Straße to Beuthener-Straße Nr. 11, I get routed all the way round the block. Using --x-no-mergeroads the routing is fine. Even stranger, when I tried to assemble a small testcase with an data-extract from JOSM, the routing is fine, even when using mergeroads. So the error seems to be triggered by something specific to the tile that was generated when building a map of Bayern. See screenshots: Bad Routing with mergeroads: <http://gis.19327.n5.nabble.com/file/n5791132/mergeroads-9452.bmp> Routing OK with --x-no-mergeroads: <http://gis.19327.n5.nabble.com/file/n5791132/no-mergeroads-19995.bmp> -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

here's the testcase: http://files.mkgmap.org.uk/download/162/mergeroads_routing_bug.tgz Test case for mergeroads routing bug containing build-script and osm Sample-Data. -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

Hi, thanks for reporting! Without having had a detailed look at your test case it sounds like the same problem reported by Gerd: http://www.mkgmap.org.uk/pipermail/mkgmap-dev/2013q4/019532.html I will have a closer look to your testcase later on to ensure that it has nothing to do with merging roads. WanMil
here's the testcase:
http://files.mkgmap.org.uk/download/162/mergeroads_routing_bug.tgz
Test case for mergeroads routing bug containing build-script and osm Sample-Data.
-- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Hi,
thanks for reporting! Without having had a detailed look at your test case it sounds like the same problem reported by Gerd: http://www.mkgmap.org.uk/pipermail/mkgmap-dev/2013q4/019532.html
I forgot to report that changing the order of nodes did not change the result in MapSource, so this is a different problem. Gerd

GerdP wrote
Hi, I forgot to report that changing the order of nodes did not change the result in MapSource, so this is a different problem.
Gerd
The case is really strange. When I build a DACH map - the routing is OK When I try a data extract from JOSM with just this part of Kaufbeuren - the routing is OK Anytime I build a BAYERN map - the routing is BAD. So only the tile generated by splitter for the BAYERN map provokes the bad routing. I suppose that the tile containing Kaufbeuren generated for the DACH map does not start at the same coordinates as the one generated for BAYERN. This is the only difference I can imagine. I uploaded the splitter-generated area information in the hope it might be helpful. http://files.mkgmap.org.uk/download/167/bayern_areas.tgz The tile in question is 65020003. Ciao, Franco -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

Hi Franco, did you check whether a tile boundary crosses the area? You can use option write-kml in splitter to create a kml file that you can load in JOSM when the opendata plugin is installed. Gerd franco_bez wrote
GerdP wrote
Hi, I forgot to report that changing the order of nodes did not change the result in MapSource, so this is a different problem.
Gerd The case is really strange.
When I build a DACH map - the routing is OK
When I try a data extract from JOSM with just this part of Kaufbeuren - the routing is OK
Anytime I build a BAYERN map - the routing is BAD.
So only the tile generated by splitter for the BAYERN map provokes the bad routing.
I suppose that the tile containing Kaufbeuren generated for the DACH map does not start at the same coordinates as the one generated for BAYERN.
This is the only difference I can imagine.
I uploaded the splitter-generated area information in the hope it might be helpful.
http://files.mkgmap.org.uk/download/167/bayern_areas.tgz
The tile in question is 65020003.
Ciao, Franco
-- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

Ahh, forget it. I remember that the data was not near the boundary. Gerd GerdP wrote
Hi Franco,
did you check whether a tile boundary crosses the area? You can use option write-kml in splitter to create a kml file that you can load in JOSM when the opendata plugin is installed.
Gerd
franco_bez wrote
GerdP wrote
Hi, I forgot to report that changing the order of nodes did not change the result in MapSource, so this is a different problem.
Gerd The case is really strange.
When I build a DACH map - the routing is OK
When I try a data extract from JOSM with just this part of Kaufbeuren - the routing is OK
Anytime I build a BAYERN map - the routing is BAD.
So only the tile generated by splitter for the BAYERN map provokes the bad routing.
I suppose that the tile containing Kaufbeuren generated for the DACH map does not start at the same coordinates as the one generated for BAYERN.
This is the only difference I can imagine.
I uploaded the splitter-generated area information in the hope it might be helpful.
http://files.mkgmap.org.uk/download/167/bayern_areas.tgz
The tile in question is 65020003.
Ciao, Franco
-- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

Hi Gerd, the KML files were generated while splitting, just forgot to put them in the archive. just in case they might be helpful http://files.mkgmap.org.uk/download/168/bayern.kml http://files.mkgmap.org.uk/download/169/dach.kml -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

The problem still persists, just switched from a bayern map to dach map. just yesterday I built a dach map with mkgmap-r2998 withe the strange routing. I could provide a new tile if this helps. -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.

Hi Franco, during the last days I found out that we probably don't write enough routing information. If I got that right, only "shorter distance" can work, but not "faster time". Garmin maps contain additional information how to get to a major road. It may well be that this lack of information makes routing worse when roads are merged. I am working on that in the high-prec-coord branch... Gerd
Date: Mon, 17 Feb 2014 00:18:06 -0800 From: franco.bez@web.de To: mkgmap-dev@lists.mkgmap.org.uk Subject: Re: [mkgmap-dev] mergeroads sometimes breaks routing
The problem still persists, just switched from a bayern map to dach map.
just yesterday I built a dach map with mkgmap-r2998 withe the strange routing.
I could provide a new tile if this helps.
-- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Hi, just to let you know: I am able to reproduce the results with your data. Not sure what happens with the mergeroad map. MapSource finds the shorter route when you select to minimize the distance, and (depending on the speed settings), it even may calculated the shorter time for the shorter route. For both routes I see a list with 7 hints, so the only explanation that I have is that MS prefers routes where don't have to change roads often within a few meters. Gerd
Date: Sun, 29 Dec 2013 10:44:24 -0800 From: franco.bez@web.de To: mkgmap-dev@lists.mkgmap.org.uk Subject: [mkgmap-dev] mergeroads sometimes breaks routing
I found a very strange routing behaviour connected to the mergeroads-patch. The Problem area is here: http://www.openstreetmap.org/#map=17/47.89307/10.63390
When I try to route ( Automobile-mode, minimize-time ) from corner Liegnitzer-Straße/Glogauer-Straße to Beuthener-Straße Nr. 11, I get routed all the way round the block.
Using --x-no-mergeroads the routing is fine.
Even stranger, when I tried to assemble a small testcase with an data-extract from JOSM, the routing is fine, even when using mergeroads.
So the error seems to be triggered by something specific to the tile that was generated when building a map of Bayern.
See screenshots:
Bad Routing with mergeroads:
<http://gis.19327.n5.nabble.com/file/n5791132/mergeroads-9452.bmp>
Routing OK with --x-no-mergeroads:
<http://gis.19327.n5.nabble.com/file/n5791132/no-mergeroads-19995.bmp>
-- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Hi Gerd and WanMil, the bug really is NOT related to mergeroads. I joined the "Beuthener Straße" and the "Allensteiner Straße" in OSM as they were split in several ways for no reason. Now the routing has changed, not for the better I'm affraid ;-) ( see screenshots ) But anyway now both mergeroads and no-mergeroads agree in how to route. I compressed the new tile with bzip2 so you have new sample data here: http://files.mkgmap.org.uk/download/163/65020003.osm.bz with mergeroads active: <http://gis.19327.n5.nabble.com/file/n5791269/mergeroads-9350.bmp> with --x-no-mergeroads: <http://gis.19327.n5.nabble.com/file/n5791269/no-mergeroads-17698.bmp> -- View this message in context: http://gis.19327.n5.nabble.com/mergeroads-sometimes-breaks-routing-tp5791132... Sent from the Mkgmap Development mailing list archive at Nabble.com.
participants (4)
-
franco_bez
-
Gerd Petermann
-
GerdP
-
WanMil