[mkgmap-dev] Should inter tile routing still work when splitting Splitter results again?

Tue Sep 29 14:28:44 BST 2009

It's a bit difficult to specify a short topic title for this, so here it 
goes:

Goal:
Render the world with tiles that are as large as possible.

Problem:
If you split e.g. America in large tiles with e.g. 1.400.000 nodes then 
you end up with a lot of tiles that do not render. If you want 
everything to render then you need to use 400.000 nodes or so which 
results in more then 1000 tiles.

You can get around this by manually defining the tiles sizes which is 
what I used to do (or use a density map, but afaik that doesn't work 
yet). So this time I've done something else:

A solution:
I created a PHP script that splits America with an high initial node 
count (of e.g. 1.400.000) and tries to render each resulting tile. If a 
tile render fails then that area is split again with an ever lower node 
count until at least two subtiles emerge which are then rendered. This 
process continues recursively until all tiles are rendered or the node 
count gets too low and splitting is aborted.

The result:
Neighboring tiles sometimes overlap because splitter seems to use a 
little larger bbox for splitting then what the source file contains. You 
can see the result of this on http://garmin.na1400.info/routable.php

The question:
Will routing be a problem because the result of recursively splitting 
and rendering results in tiles that apparently do not line up perfectly? 
I've done a few tests and it feels that routing has become worse, but I 
don't have a definitive answer. If so, what can be done about this? I'm 
curious if anyone (Chris?) has ideas about it.

I can provide the script to anyone who's interested. Just send me an 
email off-list.