Computing
Fix notation as in Section 1.5. In particular,
is an elliptic curve over
,
is an odd prime of good ordinary
reduction for
, and
is the root of
with
.
For each integer
, define a polynomial
Recall that
is the Teichmuller lift of
.
Proposition 1.20
We have that the
-adic limit of these polynomials is the
-adic
-series:
This convergence is coefficient-by-coefficient, in the sense that
if
and
, then
We now give a proof of this convergence and in doing so obtain an
upper bound for
.
For any choice
of
-th root
of unity in
,
let
be the
-valued
character of
of order
which
factors through
and sends
to
.
Note that the conductor of
is
.
For each positive integer
, let
.
Corollary 1.22
We have that
As above, let
be the
th coefficient of
the polynomial
.
Let
so that
, i.e.,
is the smallest power of
that clears the denominator.
Note that
is an integer since
.
Probably if
is irreducible then
- see
Question 1.13.
Also, for any
,
let
be the min of the valuations of the coefficients of
,
as in Lemma 1.23.
For
fixed,
goes to infinity as
grows since the
are uniformly
bounded (they are bounded by the power of
that
divides the order of the cuspidal subgroup of
).
Thus,
is a
Cauchy and Proposition 1.24 implies that that
Remark 1.25
Recall that presently there is not a single
example where we can provably show that

. Amazingly

is ``computable in practice'' because Kato has proved,
using his Euler system in

, that

by proving a divisibility
predicted by Iwasawa Theory. Thus computing elements of

gives
a provable lower bound, and approximating

using Riemann
sums gives a provable upper bound - in practice these meet.
William
2007-05-25