[maemo-developers] Maemo-mapper vector data backend follow-up

[From: Chris Lord]

On Wed, 2008-09-24 at 12:59 +0200, Till Harbaum / Lists wrote:
> Hi Simon,
> 
> Am Mittwoch 24 September 2008 schrieb Simon Pickering:
> > I fully support the work on Navit and other navigation/mapping programmes,
> > this is simply looking at an expansion of the abilities of maemo-mapper.
> I still think that what you are trying to achieve requires more work and the
> result will be less satisfying than trying to do the same with a vector program.
> 
> After have a look at the frame rates of existing navigation solutions i am pretty
> sure that there's a cpu bottleneck in rendering vector maps on these small devices. 
> The solution you are planning consumes significant more memory and cpu power 
> than the usual vector solutions (as you e.g. handle/cache bitmap tiles, redraw 
> certain things several times or render offscreen portions of tiles).
> 
> I am just afraid that a lot of man power will be consumed by such an approach
> while this would just be better invested into more promising solutions.

There are some interesting things you can do with this raster-based
approach that isn't as easy/fast with vector data though, however. A 3d
view using ray-casting, for example, would be much quicker than the same
thing using vector data, and would easily allow you to use topography
data too, to visualise height data.

Also, as maemo-mapper caches the tile data and runs asynchronously to
the tile generation, slow-down while rendering the vector data to tiles
shouldn't be too perceptible (the renderer can always be nice'd too) and
I think a raster-based mapper could run a lot faster than a vector-based
one.

There are pros and cons to raster and vector methods, but seeing as
maemo-mapper already has a substantial amount of work done, I'd
personally be interested to see this development continue :)

Regards

-- 
Chris Lord, Intel Open Source Technology Centre