|
From: Mathias <Mat...@gm...> - 2006-02-23 18:33:23
|
Hi David, On Tuesday 21 February 2006 01:04, David Luff wrote: > Thanks for all the suggestions guys. I'm ashamed to say that I'd never > heard of the kd tree or the quad-tree before. Yes, the current bucket > approach works, and yes I'll probably use it this time (Special User > Airspace records for display on the kln89 map page), but the possibility = of > a more elegant approach is something that's been bugging me for a while. It took a bit longer than expected, but here it is: A map working on a hierarchical triangular mesh like described in the other= =20 references in this thread. It could be used just like a std::map with the addition that there is a=20 range_iterator and a const_range_iterator which can be used to iterate fast= =20 over a small subset of the sphere. There is a small test program included that I used to test for correctness,= =20 that shows how that class could be used. It can stand several improovements and there are some FIXME's left. Anyway it is basically working ... But before I spend some more time with that class I wanted to konw if it is= =20 useful for your given problem and if it is really faster than what you have= =20 now. It is available from ftp://ftp.uni-duisburg.de/FlightGear/Misc_maf/SGHTMMap-20060223 Let's see where that prooves useful ... :) BTW: I am leaving tomorrow morning for the FOSEM in Brussels. Anybody else= =20 there? Greetings Mathias =2D-=20 Mathias Fr=F6hlich, email: Mat...@gm... |