[
https://issues.apache.org/jira/browse/LUCENE-8086?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16285727#comment-16285727
]
Karl Wright commented on LUCENE-8086:
-------------------------------------
[~ivera], the error distance is the linear (perpendicular) distance from a
plane to a point that is supposedly on that plane but which is not quite.
The units we're using here are not radians -- those are an angular unit.
Instead we're talking about the unit ellipsoid, where x^2/ab^2 + y^2/ab^2 +
z^2/c^2 = 1. (It is possible, as you noted before, to construct a non-unit
planetmodel, but I don't know what effects that might have, and maybe we should
put in a check to prevent it.) The error is therefore relative to 1.0, so it's
best described as a fraction of the circle or ellipsoid.
> G3d wrapper: Improve circles for non spherical planets
> ------------------------------------------------------
>
> Key: LUCENE-8086
> URL: https://issues.apache.org/jira/browse/LUCENE-8086
> Project: Lucene - Core
> Issue Type: Improvement
> Components: modules/spatial-extras
> Reporter: Ignacio Vera
>
> Hi [~dsmiley],
> The purpose of this ticket is to add a new circle shape (GeoExactCircle) for
> non-spherical planets and therefore remove the method relate from
> Geo3dCircleShape. The patch will include some simplifications on the wrapper
> and some refactoring of the tests.
> I will open shortly a pull request.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
