[Tagging] traffic lights

Bryce Nesbitt bryce2 at obviously.com
Tue Sep 6 08:55:35 BST 2011


On 09/05/2011 01:46 AM, Martin Koppenhoefer wrote:
> 2011/9/5 Bryce Nesbitt<bryce2 at obviously.com>:
>> On 09/03/2011 07:30 AM, sergio sevillano wrote:
>>
>> are we mapping reality or "for the router" ?
>> The question is at what resolution are we mapping?
>
> IMHO we should try to map at the "highest possible" resolution (our db
> has a ~1cm limit for coordinate precision, our zoom 18 is approx.
> 1:1500). Lower resolution data can generally be derived from the
> higher resolution data.
I consider also robustness, meaning "map at the highest possible level that
has a reasonable chance of receiving ongoing maintenance".

In the case of traffic lights if the intersection node is marked as 
having the proper control (e.g. none, stop light, stop sign, yield, 
TOUCAN, PELICAN, PUFFIN, Pegasus, etc), then additional detail is 
harmless.  A future simple router can route.  A future complex router 
can process the additional relations and details, when and if they gain 
sufficient traction.

Thus I encourage people not to remove the control type from the 
intersection node, but rather supplement it.  Future rendering software 
can suppress the intersection node's rendering if more detailed 
information is available.

                  -Bryce



More information about the Tagging mailing list