[mkgmap-dev] Holes in the Sea

Fri Mar 2 21:27:23 GMT 2012

> Betreff: Re: [mkgmap-dev] Holes in the Sea
> 
> The problem is that splitter and mkgmap don't work together at all. The
> solution would be that splitter does more checking against the possibility
of
> such empty tiles, or other way around - an improvement on how max-nodes
> are counted. The other possibility would be, if there was an automatic
> mechanism (mapid is important here why it needs to be somehow
> integrated) how too big tiles could be split again and reprocessed once
more
> by mkgmap.
> 

That would be perfect. But if it´s too difficult to implement, I could
imagine a simple semi-automatic approach:

Implement a mkgmap option "--seapoint-file=somefile.txt". In "somefile.txt"
the user can put a list of coordinates (points) that are somewhere in the
sea areas. If mkgmap processes a tile without coastline and there is a point
from the list inside the tile, make it sea, otherwise land. So the user can
"repair" seaholes easily by just adding one point within this area to the
text file. People may share their seapoint lists. After some time there will
be so many seapoints in the lists that holes in the sea will become
extremely unlikely.