This mysterious behavior is being done by internal
logic that detects situations of flat map projections
of spherical (presumably earth surface) data, and
removes artifacts in the display such as lines from
179 to -179 that are drawn the wrong way around. One
complexity is that in ScalarMaps like your:
RealType.Longitude -> Display.XAxis: range -180.0, 180.0
application are free to define the break anywhere they
like (i.e., other than at 180).
It is seemingly impossible to find general logic that
works correctly in all cases. One particular problem
has been very long lines like you are drawing. I recommend
breaking such long lines into short segments. Or you can
avoid this map projection logic altogether by not using
On Sun, 22 Feb 2004, Geoff Freeman wrote:
> I have been playing with EarthMaps and have met some
> behaviour of longitude which I am having trouble predicting.
> I have Scalar Maps
> RealType.Longitude -> Display.XAxis: range -180.0, 180.0
> RealType.Latitude -> Display.YAxis: range -90.0, 90.0
> This means the right edge of the map is at 180 degrees.
> I find that a line of constant latitude from -180->150 gives a line
> from -180->150. If I try -180->160 I get the "short" distance from
> Trying 10->180 on its own behaves as expected but if I include
> a 0->360 as well, the 10->180 line is drawn as -180->10 and nothing
> is drawn for the 0->360 line (I expect it is trying to draw 0->0.
> Both lines are represented as Gridded2DDoubleSet as part of a UnionSet.
> Could you give me a pointer to how RealType.Longitude is meant
> to behave. Indeed is it meant for spherical polar coordinates
> or for geographic mapping of earth data?
> Bureau of Meteorology
> Phone (03) 9669 4827