The ordinary system for naming points in the real plane
is called Cartesian coordinates (after Descartes, who invented
analytic geometry). In this system, each point of the plane has
exactly one name, consisting of a pair of numbers. (Often a
point is even thought of as being a pair of numbers, as when
we write
R for the real plane.)
You know one other system for naming points in the real
plane: polar coordinates. In polar coordinates, each point has
many names of the form
; for example, the ordinary
point
has the polar name
and also
; the origin is
for every
.
In computer graphics it is often helpful to use
homogeneous coordinates. With homogeneous coordinates, each
point of the plane has many names, each one being a triple of
numbers. You have already seen one example of homogeneous
coordinates, although we didn't call it that: In using extended
vectors for affine transformation, the ordinary name
became
.
Let's put a subscript after a triple when it
means the homogeneous coordinates of a point in the plane. That
way it won't be confused with a triple meaning a point in
R
. The subscript doesn't do anything; it's just a
reminder.
What are all the names in homogeneous coordinates of the
point whose ordinary name is ? They are simply the
non-zero scalar multiples of the extended-vector name. Some
examples of names for
are
,
,
,
,
,
, and so on.
Problem. What is the ordinary name of
?
Solution. First scale so the becomes
. Thus
describes the same point as
, which
is the point with ordinary name
.
Because one point has many names in homogeneous
coordinates, it is a good idea to distinguish between a name of a
point and the point itself. Let's write for ``the point
whose name is''. Thus in the solution to the last problem, we
could have written
. (It wouldn't really have made sense just to say the
names themselves are equal, because they aren't. This notes will
use
when it's appropriate, but you don't have to.)
Problem. What is the ordinary name of (assuming
that
)?
Solution.
.