<<<
what is the purpose of these new nodes you propose?
>>>
The purpose of these new routing nodes is to make routing data official support by osm data.
Right now, osm dos not have any properly way to model or capture routing information. So each and every routing applications have its own way to implement routing solutions.
If OSM has properly way to store routing in its database, it can provide a routing function for itself.
<<<
What is your reason for violating the first
normal-form of database-design?
>>>
The reason why i dont want routing node refer to node id to get lat and lon is because:
1. if routing node need to refer to node id to get lat lon then you need to store billion of node in memory and search though billion of node just to get the lat/lon of a routing node --> it will slow down the routing algorithmn
2. if routing node has lat/lon by itself it will be very easy to clip a set of routing node within a bound. For example: if i want to calculate a route within london city, i just need to run this ssql statement:
select * from routing_node_table
where lat and lon within london's boundary
--> this sql statement return me a set of routing node enough to do the route calculation. My application no need to consider other routing nodes outside london city (no need to consider the whole routing nodes in the planet)--> routing algorithmn run faster.
<<<
How do you suppose to generate them?
>>>
I will explain in detail how to pre-generate and/or modify routing nodes in the next email.
<<<
If you want to have a simplified data-set for
your and other's routing-applications, you are
free to render it regularly and provide access
to it for everyone who may find it usefull.
>>>
Why not standardlize routing information in one place and every routing applications refer to the same place? Right now, OSM can not calculate routing by itself.
<<<
They should not, however be a part of OSM.
>>>
Routing information should be part of OSM as a whole. The main purpose of OSM is to model the map in the real life and anything belong to the map.
If OSM does not have proper way to capture routing information then its data model is not completed.
<<<
We already have maxspeed and similar in the ways.
Travel-Times depend heavily on the type of vehicle you are
using and external factors and need to by calculated by the
routing-application.
>>>
traveltime is just an pre-estimated factor, it can be adjust by the routing application at run time by using parameter.
for example:
the pre-estimated traveltime for a highway segment route is 30mins, then depend on vehicle type parameter, the routing application will do the following:
traveltime for car = pre-estimated traveltime*1
traveltime for moto = pre-estimated traveltime*2
traveltime for walking = pre-estimated traveltime*2000 --> can not walk in highway, we set the cost of traveltime very high so the routing algorithmn will never select this segment route for walking.
<<<
Also the selection of osm-ways that are relevant for
routing depends on your vehicle. Bicycles and motor-
scooters cannot use motorways. Heavy trucks cannot
use certain bridges...
>>>
the routing application need to be flexiable enough to handle the distance and traveltime factors. By changing it at the run time, the routing applciation
can calculate relevant route for each type of vehicle.
Regards,
--- On Mon, 3/17/08, Marcus Wolschon <[EMAIL PROTECTED]> wrote:
From: Marcus Wolschon <[EMAIL PROTECTED]>
Subject: Re: [Routing] propose a new data structure for routing
To: [EMAIL PROTECTED], [email protected]
Date: Monday, March 17, 2008, 4:07 PM
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Digitalmobilemap Digitalmobilemap schrieb:
| Hi All,
|
| In order to do routing with OSM data, we need to have routing nodes
and graph
| data in osm database.
| I study OSM data structure version 0.5 but I could not find any proper
data
| structure to store routing information.
|
| I would like to propose a new data structure for capturing and/or
calculating
| routing information.
| The proposal uses xml tags similar to node,way,relation.
|
| I name the new data structure as <rtnode /> to distinguish with the
current node
| <node /> (node in osm is not
| a node, it is just a point only)
Hello,
what is the purpose of these new nodes you propose?
What is your reason for violating the first
normal-form of database-design?
How do you suppose to generate them?
If you want to have a simplified data-set for
your and other's routing-applications, you are
free to render it regularly and provide access
to it for everyone who may find it usefull.
They should not, however be a part of OSM.
| <rtnode id="757312" lat="51.3450785"
lon="-0.4606170"
| timestamp="2006-04-17T08:57:53Z">
| <rtnode ref="251490798" distance="2km"
traveltime="10minutes"
| maxspeed="60kmh" />
We already have maxspeed and similar in the ways.
Travel-Times depend heavily on the type of vehicle you are
using and external factors and need to by calculated by the
routing-application.
Also the selection of osm-ways that are relevant for
routing depends on your vehicle. Bicycles and motor-
scooters cannot use motorways. Heavy trucks cannot
use certain bridges...
Marcus
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFH3pc/f1hPnk3Z0cQRAnbJAJwKqKyEjk5we3GzlOB2QuEjsp1g5ACfWKbC
vTsRWG/0iUsZVa5ZVtGkdaI=
=Q3Gv
-----END PGP SIGNATURE----- | 