September 8th, 2010

2km in 8minutes

Can you beat that? Without using wheels :-)? And can you beat 600m in 2minutes?

And now, how do you generate such numbers from gpx track? No, gpsbabel is not tool for that job, and I could not find anything suitable. So there are now some tools in gtracks project, in the CVS.

I re-used existing "wherewasI" python script, and added custom "all_best_speeds" tool to generate "best speed on every possible distance" matrix. (Suggested usage is something like ./wherewasi.py --summary foo.gpx | ./delete_fast | ./all_best_speeds | sort -n | less). Closer look at all_best_speeds will reveal that it computes the matrix in something like O(N^3) time; I and of course the question is: can you figure out algorithm with better complexity?

And... is there some existing tool that can do this?