Re: [Gpsbabel-misc] Averaging tracks

[Kacper P. <kac...@gm...>]

Michiel, (and others too of course )

As I said before I do believe that current discussion is out of scope
of GPSBabel.
I would like implement the the algorithm in PERL, and if some want it
modern may port to PYTHON when prototype acquire production stage.
Maybe meanwhile a PARROT would acquire the production stage.

On 18 February 2011 14:35, Michiel Faber <fab...@pl...> wrote:
> Jesus Cea schreef op vr 18-02-2011 om 13:37 [+0100]:
>> [...]
>> Could you possibly describe your algorithm?.

My idea goes the same way as Jesus one. Algorithm described in
mathematical terms would help f=more then a code in particular
programming schema.

In the prior email you wrote about "simple but effective algorithm".
Below seems to be heuristic but not the algorithm that works fine in
almost any case.

> - In the set of tracks, find the one which has the moist points.
>   That wil be the reference track.

Why? I couldn't agree in such assumption.

> - For each point in the reference track, find in every other track the
>   closest point.

Definition of closest point in other tracks is another weak assumption.

>    With 3 tracks, and the the reference track has 4 points, the result will be 4 points per track.

Why just three and fourth as a reference? Appears bad.

I'll post my suggestions in the other email to not join two subjects
in one thread.

KAcper