Coordinate Systems in 2D

General coordinate systems can be viewed as parametric graphs in the same dimension.  In 2D we have a vector function of two parameters or "coordinates" such as

X[s, t] = ( x ) = (       )                                     s t                     y                                     2    2                                    s  - t

Smoothness is still the conditon

X[s + δs, t + δt] - X[s, t] = ∂X[s, t]/∂s · δs + ∂X[s, t]/∂t · δt +| (δs, δt) | · ε

where the vector  ε≈ 0  is small when the changes in input are small,  δr≈0  and  δθ≈0.  

The term  L[ds, dt] = ∂X[s, t]/∂s · ds + ∂X[s, t]/∂t · dt  is a linear parametric equation with paramters  ds and  dt.  We plot linear coordinates by drawing the vectors  ∂X[s, t]/∂s  and  ∂X[s, t]/∂t  and then filling in parallel lines in the directions of these vectors,

[Graphics:../HTMLFiles/Lect3_105.gif]

The derivative approximation says that changes in the general coordinates equal a linear term plus a term that is small compared with the change, ε≈0, when δs≈0 and δt≈0,

        X[s + δs, t + δt] - X[s, t] = L[δs, δt] +| (δs, δt) | · ε

If we magnify to make the norm of the input change, | (δs, δt) |, appear unit size the equation becomes,

        (X[s + δs, t + δt] - X[s, t])/(| (δs, δt) |) = L[δs, δt]/(| (δs, δt) |) + ε

[Graphics:../HTMLFiles/Lect3_112.gif]

Under sufficient magnification,the difference  Δ_sX = X[s + δs, t] - X[s, t]  appears the same as the vector  ∂X[s, t]/∂s · δs  and  the magnified difference  Δ_tX = X[s, t + δt] - X[s, t]  appears the same as the vector  ∂X[s, t]/∂t · δt.

Click here  to Zoom in  

[Graphics:../HTMLFiles/Lect3_118.gif]

        Unit scale        Nonnlinear mag    Linear & nonlinear mag

Figure 2

The area of the parallelogram with sides  A  and  B  is the determinant  Det[(A, B)], so the oriented area of the linear sector is

        Det[(∂X[s, t]/∂s · δs, ∂X[s, t]/∂t · δt)] = δs · δt · Det[(∂X[s, t]/∂s, ∂X[s, t]/∂t)]

The nonlinear area increment is related to this by

        δA[s, t] = Det[(∂X[s, t]/∂s, ∂X[s, t]/∂t)] · δs · δt + ε · δs · δt

where  ε≈0  when  δs≈0  and  δt≈0.  (When the microscope is powerful enough, the linear and nonlinear increments appear to coincide and we don't see  ε.)

This can all be done geometrically in the case of polar coordinates:

[Graphics:../HTMLFiles/Lect3_128.gif]


Created by Mathematica  (July 2, 2004)