Intuitively, I thought geometry would be faster than geography distance calculations, but this says otherwise: https://stackoverflow.com/questions/3547699/need-performance-on-postgis-with-geodjango#answer-3551142
The docs here also seem to suggest that distance calculations are more efficient using geography: https://docs.djangoproject.com/en/1.9/ref/contrib/gis/model-api/#srid
Why is this? Is it some sort of optimization or indexing with geography that allows faster distance calculations?
Distance calculations that assume shortest distance on the Earth will always be faster than geometry because geometry can be in any projection, and there's no way generally to know that the shortest path in that projection is the 'shortest path' and so you need to go to ellipsoid calcs and potentially insert more vertices for the curvature from the source to the target and then back transform. So yeah, do the trig for distances (unless your path is loxodromic or some other special case and you can apply shortcuts).
I'll speculate that what the GeoDjango docs are trying to say is that if you have geographic (lat/lon) data, and you want to perform range queries on that data (like ST_DWithin) in meters rather than units of degrees, then you are better served by using the geography type, which uses meters natively. The geography distance calculations themselves (ie distance(pt1, pt2)) are slower than Cartesian distance calculations.
Best explained from Microsoft in Spatial Data Types Overview:
Measurements in spatial data types
In the planar, or flat-earth, system, measurements of distances and areas are given in the same unit of measurement as coordinates. Using the geometry data type, the distance between (2, 2) and (5, 6) is 5 units, regardless of the units used.
In the ellipsoidal, or round-earth system, coordinates are given in degrees of latitude and longitude. However, lengths and areas are usually measured in meters and square meters, though the measurement may depend on the spatial reference identifier (SRID) of the geography instance. The most common unit of measurement for the geography data type is meters.
