[mkgmap-dev] Sea generation

Fri Nov 5 07:19:34 GMT 2010

On Thu, Nov 04, 2010 at 11:04:24PM +0100, WanMil wrote:
>Some loud thinking:
>Would it be helpful if we copy the Mapnik behaviour of well defined 
>coastlines? One could define one separate file that contains all 
>coastline data (from europe, from the world?). This file could be 
>maintained and improved in a better way than it is done in the OSM 
>data.  natural=coastline would be ignored/removed from the OSM input.

Good idea. This may lead to some duplicate lines, in case some 
natural=coastline double as administrative borders or something like 
that, but I am fairly sure that those cases can be sorted out.

>As a second step the separate coastline file would build a basis and 
>the data from the OSM could be compared to this. Differences in the OSM 
>files are only accepted if they don't invert the sea polygons and if 
>they don't build new gaps.

What if the coastline file were based on the database? All fixes to the 
coastline file would be backported to the database, and the coastline 
file would be re-cloned, fixed and merged back every month or so.  
Actually, we could generate a separate set of 'coastline' .img tiles 
that would be under the map data.

Should there be a simple tool for merging map layer .img files for end 
users? I am thinking that many would be happy updating their sea layer 
only once per year or so, cutting the download volume? The same could 
apply to contour layers, which might not even be distributable when 
merged with OSM-derived data.

>I haven't checked the mp warnings for long but I think most of the 
>warnings are "real" warnings now. With one exception: The conversion 
>from float to int coordinates sometimes causes some polygons to overlap 
>which don't overlap in the original OSM data.

Could the overlap check be performed in the float domain as well? If 
there is an overlap only in the int domain, then issue a 'loss of 
precision' warning? That should get rid of the bogus intersection errors 
that I see for these hyper-precision micro-mapped multipolygons:

http://www.openstreetmap.org/browse/relation/405246
http://www.openstreetmap.org/browse/relation/934321
http://www.openstreetmap.org/browse/relation/1068062
http://www.openstreetmap.org/browse/relation/1225936
http://www.openstreetmap.org/browse/relation/1232121

I haven't checked how these multipolygons look like on the map. In these 
cases, it might be OK to discard the offending polygon or the entire 
multipolygon, but what if someone starts micro-mapping boulders that are 
submerged in the beach water? Would it be always OK to discard the 
polygon that is overlapping or intersecting with the lines of other 
multipolygon members?

Best regards,

	Marko

PS: I am just updating http://www.polkupyoraily.net/osm/.