Quantcast

generate-sea: behaviour

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

generate-sea: behaviour

Ticker Berkin
Hi

I've been using --generate-sea rather than --precomp-sea for a few
weeks and there doesn't seem to be an obvious performance penalty and
the img size is slightly better (but I have been changing other things
along the way).

However - It doesn't seem to know about the rule that land must be on
the left of the coastline and sea on the right.

Taking the geofabrik download of UK/Hampshire, the Isle of Wight is
(correctly) mostly excluded - just 3 bits of coastline that don't meet
each other or the bottom edge of the tile.

With --generate-sea=multipolygon, I sometimes get the top half, which
is correctly defined land, ie has coastline to the tile edge, as sea
and the sea as nothing. The bits of disconnected coastline sometimes
have a triangle of sea where it should be, sometimes where the land is!

With extend-sea-sectors,gap=350 it is much better behaved.

Maybe the default behaviour should be to ignore these disconnected bits
of coastline (with a warning), hence both sides will show as sea.
The coast of the requested area should behave correctly.

Ticker
 
_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: generate-sea: behaviour

Gerd Petermann
Hi Ticker,

the precomp-sea option was added to solve the problems with the handling of incomplete or wrong coastline
data. I've never tried to understand the old code, so if you think you can improve the code please post
a patch.

Gerd
________________________________________
Von: mkgmap-dev <[hidden email]> im Auftrag von Ticker Berkin <[hidden email]>
Gesendet: Dienstag, 14. März 2017 11:20:40
An: mkgmap development
Betreff: [mkgmap-dev] generate-sea: behaviour

Hi

I've been using --generate-sea rather than --precomp-sea for a few
weeks and there doesn't seem to be an obvious performance penalty and
the img size is slightly better (but I have been changing other things
along the way).

However - It doesn't seem to know about the rule that land must be on
the left of the coastline and sea on the right.

Taking the geofabrik download of UK/Hampshire, the Isle of Wight is
(correctly) mostly excluded - just 3 bits of coastline that don't meet
each other or the bottom edge of the tile.

With --generate-sea=multipolygon, I sometimes get the top half, which
is correctly defined land, ie has coastline to the tile edge, as sea
and the sea as nothing. The bits of disconnected coastline sometimes
have a triangle of sea where it should be, sometimes where the land is!

With extend-sea-sectors,gap=350 it is much better behaved.

Maybe the default behaviour should be to ignore these disconnected bits
of coastline (with a warning), hence both sides will show as sea.
The coast of the requested area should behave correctly.

Ticker

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
Loading...