It would be interesting to see if the same system could be implemented for English. Say, make a text renderer that groups common sequences of 2-to-4 letters into one distinct (but still immediately legible) glyph; the set of available glyphs would remain constant. With practice, readers' brains might learn the glyphs as a "shortcut" to take in these sets of letters- almost like a very basic compression scheme programmed into your reading-comprehension neurons.