We use random context picture grammars to generate pictures through success
ive refinement. At any stage a picture consists of a shape divided into sma
ller shapes, each containing a variable or terminal. A variable may be rewr
itten according to a production of the underlying grammar, which entails ei
ther dividing the shape containing it into smaller shapes, or substituting
a variable or terminal for it. A production may depend on context randomly
distributed in the intermediate picture. Context is classified as either pe
rmitting or forbidding, the former enabling the application of a production
, the latter inhibiting it. For visualization purposes every terminal is as
sociated with a color, and its shape filled with that color.
We show examples of pictures generated with random context picture grammars
. Then we concentrate on grammars which use permitting context only and pre
sent a pumping lemma for the corresponding picture sets.