It
seems that a lot of early character displays were either stroke
generated, or used charactron(sp?) tubes (electron beam first directed
through a mask with character stencils, then deflected to the correct
position on the screen), with raster coming later.
There were competing raster based terminal systems in the 1960s.
The IBM approach was to have one big box do all the thinking for all
the terminals, with the terminals being not much more than TVs with
keyboards (2848/2260) - I suppose with one brain doing all the work,
you could save a bunch of expensive circuitry.
The other approach was to use a tiny monoscope, with the target having
an array of characters that could be raster scanned by a jumping beam.
This is not as kludgey as it sounds.
--
Will