Observing the rise of Alacritty and Kitty, I had to concede that hardware-accelerated terminal emulators had clearly become a thing. As a heavy user of the humble text terminal, I am naturally interested in such fresh perspectives on this several decades old UNIX-ism. Somewhat surprisingly, after trying both of these programs, none of them turned out to be the technological facelift I was looking for – and none of them work on my low-spec single-board computer, since they require “desktop” OpenGL while my ARM Mali GPU only supports OpenGL ES. The critique of these programs is for a later post; here I want to talk about something more fundamental.
If you would set out to create the theoretically most efficient way of rendering the content of a text terminal to a window image, how would you go about it?
That is, how would you transform a grid of text (a two-dimensional array of characters, possibly enhanced with attributes such as color) into a pixel image?