-
Rn is complete
with usual metrics (Cauchy sequences converge).
A sequence converge if d(xi , limit) -> 0 on your surface.
A Cauchy sequence if such that d(xi , xj ) -> 0 when
i and j goes to infinity but they may not converge to a limit. In a complete
space only they do and any Cauchy sequence converge. Counter-example: an
open set in Rn , with a sequence
that goes infinitely close to the "boundary". We will consider the space
as coomplete after this point.
For fanatics, completeness give very useful results with series, see
link.
In a complete space, Hausdorff distance and fixed point theorem are
the basis of Fractal theory, which lead for instance to fractal image compression
or to Julia sets.
-
Continuity
You don't need to be Super Mario to navigate on your surface.
If a sequence converges, then it converges also on your surface: (xi
) cv => f(xi ) cv
---->
-
Differentiability
We can define a tangent at every point: a line in 2D, a plane in 3D,
a hyperplane in higher dimensions.
For Implicit functions, the normal vector is defined by the gradient:
For Explicit functions, the normal vector is, for a single linear map:
-
Related operators: Div, grad, curl
and all that... thanks to read the book by Schey.
-
C0, C1, C2, Cinf:
higher degrees of differentiability
Need to consider differentials as linear functions and iterate the
processus. Tied to differentiability of partial derivatives, even if there
is some subtilities you have to watch. See reference books.
Final remark: