Running this in the Talon REPL shows that .dpi is the hypotenuse of .dpi_x and .dpi_y:
from talon import screen
import math
s = screen.containing(0, 0)
s.dpi_x
s.dpi_y
s.dpi
h = math.sqrt(s.dpi_x ** 2 + s.dpi_y ** 2); h
abs(h - s.dpi) < 0.000001
(h - s.dpi) == 0
Although even this is incorrect, because the DPI of the 45° angle would be smaller, not greater than the x or y DPI.
This is misleading, because .dpi is perceived as a tolerable substitute of .dpi_x/.dpi_y. It would be more sensible to have the average of .dpi_x and .dpi_y in .dpi.
Running this in the Talon REPL shows that
.dpiis the hypotenuse of.dpi_xand.dpi_y:Although even this is incorrect, because the DPI of the 45° angle would be smaller, not greater than the x or y DPI.
This is misleading, because
.dpiis perceived as a tolerable substitute of.dpi_x/.dpi_y. It would be more sensible to have the average of.dpi_xand.dpi_yin.dpi.