CS488 - Introduction to Computer Graphics - Lecture 4

Graphics I/O Hardware

Mice & Keyboards (Input)

Types of Input

Request
Sample
Event

Devices

Text Input
Pointing Input

Event-based Programming

Event Loop
- Xlib
Callbacks
- Toolkit programming
- Java listeners
Callbacks need formal work. Java is trying hard but needs to step further back from the coal face.

Geometry

The course notes give you the axioms and theorems. In class, I will concentrate on ideas and algorithms.

Euclid and the Greeks

Of course, it's completely unfair to give the Greeks all the credit.

Straightedge and compass.
Congruence: two different triangles share a property called congruence.
Similarity: two different triangles share a propety called similarity.
Felix Klein's Erlangen program. (circa 1870).

Descartes

Points are pairs (triples, etc)

Connects geometry to algebra.
1. Subtract two points
2. Add (with a little hesitation) two points
3. Find a additive unit, the origin, and an additive inverse
4. Multiply a point by point???
  - perhaps as complex numbers (in 2D at least)
  - complex numbers don't generalize to 3D
  Multiply a point by a scalar
5. Find a multiplicative unit, the scalar 1.
6. Does a multiplicative inverse exist?
Addition/subtraction might work all right, but the prominent role played by the origin is suspect. Why?
Computers do algebra, not geometry.
Distance is defined: more strongly than the compass!

Points morphed into vectors. What is the difference between two points?

An arrow that goes from one point to the other
The arrow does not change when the origin is translated
Suppose we think only about vectors
1. Subtract
2. Add
3. Unit, inverse
4. Scalar multiply
  Closed up to here
5. Dot product of two vectors: defines length, angle.

Makes geometry challenging as an interpretation of algebra!

Vector Spaces, aka Linear Algebra.

We have some things that look like Descartes' points, though we don't know exactly what they are.

We can tell when they are equal

v1 =1 v2 if and only if x1 = x2 and y1 = y2.

We can add them together

v1 +1 v2 means (x1, y1) +1 (x2, y2) =1 (x1 + x2, y1 + y2)
They are closed under addition.
There is a zero element 01 =1 (0, 0)
There is an inverse: If x1 =1 x2 and y1 =1 y2, then (x1, y1 ) +1 (x2, y2 ) =1 01

Where addition is, multiplication cannot be far behind

a *1 v =1(a * x, a * y ), called scalar multiplication.
They are closed under scalar multiplication.
There is a unit element 1; there is a zero element 0.
Scalars must be a field.
Scalar multiplication distributes over addition: a *1 ( v1 +1 v2 ) =1 a *1 v1 +1 a *1 v2.

What are these things? We want the geometrical interpretation.

Descartes called them points, BUT they are not like points.

How?

You can move them around, like
- derivatives
- position independent code.

We call them vectors.

Generators

How few vectors can I use to create all of them?
The number is the dimension.
A set having that number, with appropriate weasel words, is a basis.

Transformations

Vectors can be transformed, by multiplying them by matrices

Rigid transformations in 2D
1. Unit
2. Rotation
3. Reflection
More complex in 3D
Non-rigid transformations in 2D
1. Scale
2. Skew

If you are Felix Klein you show that they are a group, and how they separate into subgroups. But even though you are not Felix Klein your response to this formulation is the same.

What about translation?

Affine Spaces

Intuition is that when I add an origin, O, vectors define points.

Each point, P, can be - but is not - identified with the vector that joins it to the origin.
- P =2 O +2 v
- This is point/vector addition.
- =2 and +2 are related to =1 as follows,
  - First, P1 =2 O +2 v1 =2 P2 =2 O +2 v2 if and only if v1 =1 v2.
  - Second, ( O +2 v1 ) +2 v2 =2 O +2 ( v1 +1 v2 )
Define point subtraction
- v =1 P1 -2 P2 if and only if P1 =2 P2 +2 v