|
| 1 | +!!! danger |
| 2 | + |
| 3 | + You are currently reading a DRAFT that is available publicly to facilitate collaboration |
| 4 | + |
| 5 | +<script src="../../js/tutorial.js"></script> |
| 6 | + |
| 7 | +This guide is meant to be a gentle intro to Projective Geometric Algebra, |
| 8 | +also referred to as $\mathbf{P}(\mathbb{R}^*_{3, 0, 1})$. It assumes no |
| 9 | +knowledge of quaternions and dual-quaternions, but if you have some familiarity |
| 10 | +of either, you may find yourself armed with a new appreciation for them. Also not assumed |
| 11 | +is any knowledge of abstract algebra, or any specific algebra closely related to Geometric |
| 12 | +Algebra (e.g. exterior algebra, Clifford Algebra, etc.). |
| 13 | + |
| 14 | +The emphasis first is on just getting familiarity with the notation and the various |
| 15 | +operations and what they do. The [references](../../references) page on the left contains |
| 16 | +some excellent references if you prefer a bottom up approach. Here though, the goal will |
| 17 | +be to build your intuition primarily through examples, and then introduce the formalism |
| 18 | +afterwards. |
| 19 | + |
| 20 | +!!! tip |
| 21 | + |
| 22 | + Grab a pen and paper! You are expected to work out a number of expressions by hand |
| 23 | + and _see_ for yourself that the formulae behave as advertised. Further, drawing pictures |
| 24 | + is _very_ important to maintain the linkage between the algebra and the geometry. |
| 25 | + |
| 26 | +We're going to go straight to 3D, so hang on tight. Let's start with three perfectly |
| 27 | +ordinary basis vectors, $\ee_1$, $\ee_2$ and $\ee_3$. Now, normally when we think about |
| 28 | +vectors, we imagine that they have some length and direction. In this case, let's have |
| 29 | +$\ee_1$ point in the x-direction, $\ee_2$ point in the y-direction, $\ee_3$ point in the |
| 30 | +z-direction, and give all of them unit length. Each one of these basis vectors can be |
| 31 | +scaled by a weight, and we can take linear combinations of them to create any vector |
| 32 | +in our 3D space. So far, everything behaves just like your good ol' 3D vector space. |
| 33 | + |
| 34 | +Now, let's pause and consider for a moment what our vector space might be lacking. With |
| 35 | +vectors alone, we can certainly come up with ways to represent all sorts of things. |
| 36 | +Sometimes, vectors are arrows from the origin. Other times, we use vectors to mean the |
| 37 | +point terminated at by that arrow. Still other times, a vector is used to represent a |
| 38 | +plane through the origin by encoding the normal to the plane. In a way, vectors are |
| 39 | +somewhat encumbered due to the need to represent _all_ the various entities in geometry |
| 40 | +one way or another. But even if we try, we'll find that there are still aspects of geometry |
| 41 | +that we can't reasonably or easily represent with vectors. For example, what if the |
| 42 | +plane didn't go through the origin? I suppose we could use two vectors, one for the normal, |
| 43 | +and one to describe a point in the plane. What about a rotation? Maybe we use a vector |
| 44 | +for the axis, and a number for the rotation. Translations? Maybe the vector points in the |
| 45 | +direction of the translation and the length encodes the displacement. Do you see |
| 46 | +an issue with the way things are going with this thought exercise? All of these |
| 47 | +interpretations of what a "vector" is are not mutually compatible with one another! We |
| 48 | +certainly can't add a vector intended to be used as a rotation axis and a vector intended |
| 49 | +to be used as a plane normal and expect to have a consistent interpretation of the result. |
| 50 | +All of them need to be treated distinctly and live "in their own space" as it were, with |
| 51 | +very delicate code to keep the invisible boundaries between them uncrossed. |
| 52 | + |
| 53 | +Needless to say, mathematically, the situation described above leaves much to be desired. |
| 54 | +What we'd like is an algebra (aka the Geometric Algebra) that could describe all the entities |
| 55 | +we need (points, lines, planes, rotations, translations, to name a few) in a _unifying_ |
| 56 | +framework glued by an operation which has a sensible meaning when its operands are any |
| 57 | +of the listed entities (aka the geometric product). To make this a reality though, we're |
| 58 | +going to need to move past our vector space and limited set of operations to an algebra |
| 59 | +that is much richer. This algebra (the Geometric Algebra) will have more operations than |
| 60 | +you're used to, and more "things" than you're used to, but that's to be expected. After all, |
| 61 | +geometry is far richer than just arrows emanating from the origin. Embracing the additional |
| 62 | +structure is in some sense, akin to embracing the reality that is geometry itself, and |
| 63 | +acknowledge that the algebra is going to need to play some catch up to describe the |
| 64 | +geometry more aptly. So, then, aside from vectors, what else do we need? In the next |
| 65 | +section, we'll describe the exterior algebra as a stepping stone to geometric algebra, |
| 66 | +so see you there! |
0 commit comments