Wheather Routing in 1.95

[Captain nono]

Hello,

I'm trying using WR for a while, playing with parameters to understand their interest and benefits, but I cannot success in getting the true Best Route with the tool (meaning better the ones I can draw with the Route tool).

I found out 2 problems :
1. the Max angle parameter seems not to be used because it's impossible to get a track at 60° from destination heading. Whatever number of new segments you set, the isochrones never reach the 60° limit
2. it seems that the tool generate new segments only on one side (tack) of direct destination, i.e. on the most rapproaching tack.
or, more probably, the tool prunes non-rapproaching tracks on priority one, even with wake angle close to 0.
Final outcome is that Best Route starting with a non-rapproaching segment are never kept nor selected as best ones

3. no matter your responses to 2 first points, I suggest (if not the case) that the new segments are calculated on both sides of direct course (of course) and using an heading increment of (Max angle)/(Nb new segments) / 2

Last, I am used with other routing software and your tool seems pretty slow. I am quite sure (as a real sailor and a real IT programming expert) that your algorithm could be improved : if you like I will be keen (and proud) to contribute to your freeware by some optimisation advice (or work). Post me a private mail if it's the case, Omar.

[omar]

The WR calculator "branching algorithm" goes like this:
1- for a <given point> calculate the direct course to destination.
2- using the vertices of the polar chart, find the course closest to
the direct course. Add a segment in that direction.
3- Add segments alternately to each side of the direct course
until the "max branches" is reached. ( not a single tack )

For example, if the polar chart has points each 5 degrees
and you are using 15 branches, you will have 7 branches
to each side. That would cover a 35 deg angle to each side
of the direct course.

You may ask: How about upwind sailing? Shouldn't it test at least 45 deg to each side?
It does. Since polars (normally) don't have points in the 0-30 deg range,
these angles are skipped, into the other tack. For upwind directions,
15 branches cover some 50°.

As it is, the algorithm assumes that the polar chart database has some
particular format ( vertices separation=5°/ TWA from 30° to 180°).
Other formats should work too, but the three size will grow fast
for small separations. That is why the "full polar databases" is not recommended.

But you are right when you say that "[wind angle] never reach the 60° limit"

As for speed, the only other software I tried is Zezo's http://www.zezo.org/
which seems pretty fast. But I did not run any speed test.