Hi all, I see this completly different : many people use osm data for routing. Routing is the main purpose for using OSM. If routing [ quote] : it will look a bit strange [end quote] is not akzeptable.

Thomas

Quote :.............

So yes with very high DP values and lots of simplification if you plan a car route over highways, at some point it will look a bit strange. But again, very few people use Garmin maps for car routing nowadays, and even less car routing based on OSM data. ......

On Wed, 5 May 2021 at 14:29, Gerd Petermann <gpetermann_muenchen@hotmail.com> wrote:

Hi Felix,

I don't think that it only depends on the typ file. Programs like GpsMapEdit also show wrong merges.

Reg. roads: Seems RoadMerger also doesn't reverse points, I didn't notice that before. So, I have two methods to improve.
It's quite difficult to understand the effects because we
- merge roads with equal attributes
- possibly split those roads again to avoid loops or too long roads or too complex roads (too many connections) or other limits in IMG format (this is needed for routable maps, maybe not for others)
- possibly merge again at levels > 0

I have to think about this again. Maybe there is no problem with the NET file changes when this last step is done. The NET file contains "pointers" to the RGN file for each level in which the road is rendered. So, if a road is rendered at level 1 the data in NET allows to show all the road labels. Maybe that's all and there is no further effect (esp. not on routing), but maybe the info is also used when you create route in Basecamp. I guess it's not.
I have no idea what Garmin software does when a road has pointers to lines at levels 0, 1, and 3 but not at level 2. I found no such case in Garmin demo maps. I think that could happen when I don't add code to prevent it.
Probably it doesn't matter much when RoadMerger is allowed to reverse so that we have fewer parts which could be merged later.

Gerd