On Thu, Apr 3, 2008 at 5:25 PM, Brett Henderson <[EMAIL PROTECTED]> wrote:
> By default, osmosis will truncate at the last node inside the boundary > which isn't always ideal. The completeWays argument will include all > way nodes outside the boundary which again isn't perfect. > > The new PostGIS based bounding box functionality will never truncate > ways but may leave out nodes outside the box depending on whether > completeWays is specified. > > Introducing truncation with phantom nodes is very difficult to do in a > memory efficient manner. If an in-memory implementation was sufficient > it would potentially be much simpler but doesn't exist yet. The new > osmosis PostGIS schema would be a good starting point for this. A > single task implementation may need to be used (ie. not one bbox task > instance per tile) so that it can maintain lists of phantom nodes > created at boundaries shared between multiple tiles. > > I think it was Karl who was looking at creating a tiling task. Not sure > whether he's still looking at this. I'd like to play with this but > realistically I'm not going to find the time any time soon. Yep, I'm still planning to do it, but I've just lost my job and surprisingly have even LESS time to work on OSM stuff. It's all going to be tied in with my new proposed rules file (so I can deal differently with the variety of ways--routable, polygons, and simple polylines), so if anyone wants to take a look at it and comment about the structure, etc., it's on my Wiki page here: http://wiki.openstreetmap.org/index.php/User:SiliconFiend/Ruleset I need to do some proper OOA&D for my addition, too, and I want to play with the UML editor for Eclipse. But, along with my job I lost my great development laptop, so I'm trying to get back up and running on my older, slow laptop with limited hard drive space. Karl
_______________________________________________ talk mailing list talk@openstreetmap.org http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
