Cryptography Reference
In-Depth Information
that is defined for at least one point in E ( K ) (so, for example, the rational
fu nction 1 / ( y 2
x 3
Ax
B ) is not allowed). The function takes values in
K
.
There is a technicality that is probably best described by an example. Sup-
pose y 2 = x 3
∪{∞}
− x is the equation of the elliptic curve. The function
f ( x, y )= x
y
is not defined at (0 , 0). However, on E ,
x
y
y
=
1 ,
x 2
which is defined and takes on the value 0 at (0 , 0). Similarly, the function y/x
can be changed to ( x 2
1) /y , which takes on the value at (0 , 0). It can
be shown that a function can always be transformed in this manner so as to
obtain a n expression that is not 0 / 0 and hence gives a uniquely determined
value in K
.
A function is said to have a zero at a point P if it takes the value 0 at P ,
and it has a pole at P if it takes the value
∪{∞}
at P . However, we need more
refined information, namely the order of the zero or pole. Let P be a point.
It can be shown that there is a function u P , called a uniformizer at P ,with
u ( P ) = 0 and such that every function f ( x, y )canbewrittenintheform
f = u P g,
with r
Z and g ( P )
=0 ,
.
Define the order of f at P by
ord P ( f )= r.
Example 11.1
On y 2 = x 3
x , it can be shown that the function y is a uniformizer at (0 , 0).
We have
1
x = y 2
1 ,
x 2
and 1 / ( x 2
1) is nonzero and finite at (0 , 0). Therefore,
ord (0 , 0) ( x )=2 ,
and
ord (0 , 0) ( x/y )=1 .
This latter fact agrees with the above computation that showed that x/y
vanishes at (0 , 0).
Example 11.2
At any finite point P =( x 0 ,y 0 ) on an elliptic curve, the uniformizer u P can
be taken from the equation of a line that passes through P but is not tangent
 
Search WWH ::




Custom Search