[Routing] propose a new data structure for routing

[Marcus Wolschon]

> OSM is a database where we collect data independend of it's purpose. --> The routing information that I propose is independend to every routing-applications.
> 
> The relation between a street to routing nodes is independend to every routing-application, isn't it? I am talking about the routing nodes at every junctions of two or more streets. The relation between a street to routing nodes is always there until the street is changed (closed or extended) so this relation is independend to routing-application. It is real-life data of street map.
> 
> The distance between two routing nodes is also independend to every routing-applications, isn't is? Unless the street is changed, the distance between 2 routing nodes will not be changed.


Actually this is not independent of the routing-application.
there are many metrics for "distance".
* 2D length of ways
* 3D length of ways
* estimated travel-time for vehicle A by algorithm X
* estimated travel-time for vehicle B by algorithm X
* estimated travel-time for vehicle B by algorithm Y
* fuel-consumption for vehicle A by algorithm X
* scenic value for history-tour
* scenic value for photo-tour
* ...

> but it captures all the data needed for routing too --> I don't think the current OSM data-format of it's API is capturing all the data needed for routing.
> for example, we don't know which node id (point id) of a way is a routing node

All you need is to look at the "highway"-tag here and use the ways that
are legal to route over for the vehicle-type you are routing for.

 > we don't know which node id is a junction to other ways unless we 
scan all the ways in OSM database to search for the ways that has the 
same node id.

A simple "fetch all ways in bounding-box" is enough for a
routing-application here. A program reads OSM-data and
builds it's own data-structures with only the information
it requires.

> I thought this routing email list is for building internal routing solution for OSM. But it is not. This routing email list is for building routing solution outside OSM's core, isn't it.

That is correct. The list of for the many people who program
their routing- and navigation -solutions using OSM-data.


PS:
Your quoting-style is very hard to read.

Marcu