Suppose that a man is sitting in front of a computer calculating a func￾tion over its domain. In its simplest realization the program keeps printing
out the samples of the domain with their respective function value in al￾phanumeric form. The user of the program who is mainly interested in
the shape of the function has a very hard time reading all the data before
they are scrolled o the screen, interpreting numbers like 1:2345e12, and
constructing the shape in his mind. He would prefer the computer to cre￾ate the drawing of the function itself, and not to bother him with a list of
mind-boggling
oating point numbers. Assume that his dream comes true
immediately, and the computer draws horizontal rectangles proportional to
the function value, instead of printing them out in numeric form, making
a histogram-like picture moving up as new values are generated. The user
can now see the shape of a portion of the function. But is he satised? No.
He also wants to have a look at the shape of larger parts of the function;
he is not happy with the reduced accuracy caused by the limited resolution
of the computer screen, meaning for example that two values that are very
close to each other would share the same rectangle, and very large values
generate rectangles that run o the screen. It irritates him that if he turns
his head for just a second, he loses a great portion of the function, because
it has already been scrolled o. The application of graphics instead of a
numeric display has not solved many problems.