[OSM-talk] About large streets

[Collinson Mike]

Reading your email it occurred to me that the O'Reilly book "AI for Games 
Developers (Creating Intelligent Behaviour in Games)", Bourg and Seemann, 
might provide some interesting insights.  I shall certainly re-read it to 
see if it is relevant.  The first 7 (of 15) chapters are basically about 
path finding.

Mike Collinson

At 09:30 AM 23/02/2006 +0000, Simon Hewison wrote:
>That was in the days of a 1MHz cpu, and relatively little RAM. These days, 
>most PC-based route planning software seems to do things differently. I 
>found a useful link.
>
>http://ai-depot.com/BotNavigation/Path.html
>
>So, when once of us (possibly me) starts a route-planning bit of software 
>using OSM data, it would first of all need to get a data snapshot of an 
>bounding box a little bigger than the start and end points, and then build 
>it's own data structure, knowing that if there's a segment from A->B then 
>unless there's tags otherwise, B->A is also possible. Therefore the OSM 
>database doesn't itself need to store segments from A->B and B->A because 
>that could be left to the route planning engine's own data structure.