Measuring on the real Earth

Debrief v4 measures ranges and bearings along the true shape of the Earth. Legacy Debrief did not, so some figures will change slightly.

Legacy Debrief treated each small patch of sea as flat, with one degree of arc equal to 60 nautical miles. Debrief v4 measures along geodesics on the WGS84 ellipsoid, the same model as GPS, whatever projection the map uses.

In legacy's code, range and bearing (WorldLocation.rangeFrom, bearingFrom and subtract) go through its FlatEarth model. That is mid-latitude sailing: the change in longitude is scaled by the cosine of the mean latitude, and the range in degrees is turned into nautical miles at 60 to the degree (Conversions.Degs2Nm). The desktop application did install a GeoTools calculator, GeoCalculatorAdapter, which is ellipsoidal, but only its bearingBetween method uses it. Everything else, including every range, still used the flat calculation it inherits.

The difference is small but not zero. Worst cases, across every bearing:

Latitude	Range	Range differs by	Bearing differs by
50°	1 nmi	7 m	0.09°
50°	10 nmi	70 m	0.16°
50°	50 nmi	350 m	0.5°
70°	50 nmi	430 m	1.1°

flowchart LR
    old["Legacy: flat patches<br/>1° = 60 nmi"] -- "about 0.4% apart<br/>at UK latitudes" --> new["v4: WGS84 geodesics"]

Try it: drag A and B and watch the two answers part company. Move them further north, or further apart, and the difference grows. The dashed line is just a straight line on the Mercator chart, which is how neither of them measures.

If you compare a v4 figure with one from an old report, expect differences of this size. Bearing residuals in target motion analysis are the most sensitive: a solution built in legacy Debrief will not show exactly the same residuals in v4.

Every value shown on screen also goes through one formatting service, so knots, nautical miles, three-figure bearings such as 045.0°T, and Zulu times look the same everywhere.