20 Oct
2008
20 Oct
'08
12:41 p.m.
They actually
scan to the upper left corner of each character, then
sweep out the 5x7 dot matrix for that character, before adjusting the
x,y deflection to the upper left corner of the next character.
I seem to recall the DEC VT11 (and thus GT40) did something similar. When
it displayed a character it scanned a little raster at the appropriate
place on the screen and used a normal dot-matrix character generator to
turn the beam on and off.