Or the data is out of date and the interchange has been reworked; I’ve seen both. But it’s also likely that someone entered the data before there was good aerial imagery of the spot and was going from memory and estimates, which is not horrible, since it’s right enough to be better than no data (even if the topology is a little off).
As always: also note that the bing aerial imagery is frequently misaligned. That could explain the off-ness of the roads in the north (usually those roads are not moved a tiny distance during road rebuild). Of course it also could be the other way round and the old aerial imagery which was used to trace was misaligned. Look for source tags and the date of creation.
If you look at the Bing imagery a bit closer on the northern part of the traffic circle you can see an arc of darker green grass, my guess it the traffic circle might have been more round in the past and they reconfigured it in the past few years.
Also note the line of browner grass on the old route of the primary highway, this further leads me to believe it was this way in the past.
The main thing wrong with this I would say is the way the northern connections to the roundabout are made. The “on” flares share the same nodes as the “off” flares. A router will think it’s possible to traverse from the road on the east to the road on the north without traversing the roundabout at all.
A router generating a list of directions won’t list this as a “take first exit on roundabout”. It may even decide not to generate a turn instruction if many of the two ways tags are the same.
This is because the highway junction was traced according to the old junction shape in the former version. Look here how it was before, seen on GE:
The mapper before you made a good job in terms of the junction shape, is not the case that he is not able to trace a road. Roads may change, and the mapper before you simply did still not have the updated imagery. I agree with you that the mixture of highway tags in that junction was not necessary.
OpenStreetMap
Discussion
Kommentaar van !i! op 28 Oktober 2012 om 06:57
Maybe BING is out of date?
Kommentaar van asciipip op 28 Oktober 2012 om 13:05
Or the data is out of date and the interchange has been reworked; I’ve seen both. But it’s also likely that someone entered the data before there was good aerial imagery of the spot and was going from memory and estimates, which is not horrible, since it’s right enough to be better than no data (even if the topology is a little off).
Kommentaar van Sanderd17 op 28 Oktober 2012 om 13:50
The only thing that really bothers me is the classification. Why have a motorway or motorway_link between a secondary and tertiary road?
Kommentaar van aseerel4c26 op 28 Oktober 2012 om 20:24
As always: also note that the bing aerial imagery is frequently misaligned. That could explain the off-ness of the roads in the north (usually those roads are not moved a tiny distance during road rebuild). Of course it also could be the other way round and the old aerial imagery which was used to trace was misaligned. Look for source tags and the date of creation.
Kommentaar van Sundance op 28 Oktober 2012 om 21:05
If you look at the Bing imagery a bit closer on the northern part of the traffic circle you can see an arc of darker green grass, my guess it the traffic circle might have been more round in the past and they reconfigured it in the past few years.
Also note the line of browner grass on the old route of the primary highway, this further leads me to believe it was this way in the past.
Kommentaar van robert op 28 Oktober 2012 om 22:42
The main thing wrong with this I would say is the way the northern connections to the roundabout are made. The “on” flares share the same nodes as the “off” flares. A router will think it’s possible to traverse from the road on the east to the road on the north without traversing the roundabout at all.
A router generating a list of directions won’t list this as a “take first exit on roundabout”. It may even decide not to generate a turn instruction if many of the two ways tags are the same.
Kommentaar van leodobrasil op 28 Oktober 2012 om 23:21
This is because the highway junction was traced according to the old junction shape in the former version. Look here how it was before, seen on GE:
The mapper before you made a good job in terms of the junction shape, is not the case that he is not able to trace a road. Roads may change, and the mapper before you simply did still not have the updated imagery. I agree with you that the mixture of highway tags in that junction was not necessary.
Thank you for updating that junction. Leodobrasil