[OSM-talk] mkgmap confusion: zoom, resolution and levels.

[Steve Ratcliffe]

On Thu, Sep 11, 2008 at 12:04:54PM -0400, simon at mungewell.org wrote:
> The map-features file contains a value for the minimum resolution value at
> which a feature is present.
> 
> However there is also the levels concept; which is what I'm having trouble
> getting my head around. Where the downloadable map contains different
> levels, which are rendered according to 'zoom' state (or something like
> that - what do I know... I'm confussed).

A garmin map has several separate maps that are prepared at different
levels of detail.  These are the 'levels'.  You might have 0, 1 and 2.
The map maker can then choose the resolution for each of the levels.
If you associate level 0 with resolution 23 and level 1 with resolution 20
for example, then that determines the zoom at which you switch from the
level 1 map to the level 0 map.  If you made the level 0 map at resolution
24, you would have to zoom in further before the device switched to it.

> Does this mean that the steps (zoom state where things become visible) are
> limited to the number of levels, or are the minimal resolutions respected
> despite the number of levels.
> 
> If the later, what is the point of having the levels concept?

The only reason why the map-features file specifies by resolution is so
that you can change the level to resolution association while still getting
reasonable results.  If you are working from scratch then you can just
decide level and resolution and only place features at one of the chosen
resolutions.

> Example:
> If I have feature1 at 18m, feature2 at 20 and levels set to
> 0=24,1=22,2=16,3=14. Will feature1 and feature2 first appear at the same
> zoom state or at different zoom states?

Feature1 will be in both the level 1 and level 0 maps, while feature2 will be
just in the level 0 (most detailed) map.

However there is a further wrinkle, just because something is in the level
1 map doesn't mean that the device will actually display it.  This is
particularly true for POI's, the device will just refuse to display
something if it doesn't want to.  To see restaurant etc you often have to
zoom in even further than you would expect - more than when the level 0 map
first becomes visible.  Roads are pretty consistant though.

> Lastly (on the assumption that it saves space or speeds up rendering) is
> there a way to reduce the complexity of a feature for lower resolutions?

Yes, it is a good idea to reduce the complexity of the features at lower
resolutions and mkgmap does do some automatic reduction, but it is fair to
say that it could be better.

..Steve