#12 Finding nearest centerline fails

[Xavier]

Finding the nearest centerline is failing numerous times.
For example for TSE 2017, the distance displayed from Paris should be about 5,020 km, however Sciatrope doesn't display any value! When closer to the eclipse path the distance and heading are displayed, nevertheless the given values do not always appear to accurate.

Another remark. In your Sciatrope description you say
"Eclipse circumstances are displayed in tabular form, on a map, a compass, or in an image, with azimuthal or equatorial orientation, which can be flipped horizontally, vertically, or both."
However in the latest 0.2 release I don't see the describe features. Where are they?

[Jean-Etienne LAMIAUD]

The point is calculated by finding the intersection of the eclipse cone axis with the observer's local plane, the axis speed vector in this plane, finding the time when the segment "observer to eclipse axis intersection point" and speed vector are orthogonal, using the eclipse axis intersection point at this time, and then iterating.

A better algorithm would be (starting at eclipse maximum time):

• getting the umbra speed vector in the centerline point plane (tangent to the geoid)
  • by projecting the umbra speed vector in the local plane, along the umbra axis
  • and then adding the earth rotation speed vector in the local plane
• calculating the time distance between the plane, orthogonal to this speed vector originating on the centerline point, and the observer.
• iterating again to this new time

If there is no maximum eclipse for the observer, start at eclipse maximum time (jd in Bessel elements).