---
title: Twisted Elliptic Curve Codes (TECCs)
url: https://www.emergentmind.com/topics/twisted-elliptic-curve-codes-teccs
type: topic
---

# Twisted Elliptic Curve Codes (TECCs)

Searching arXiv for the specified paper to ground the article in the cited source.
Twisted elliptic curve codes (TECCs) are a class of algebraic–geometric codes introduced as a twisted analogue of elliptic curve codes, motivated by the studies of twisted generalized Reed-Solomon (TGRS) codes. The paper "On a class of twisted elliptic curve codes" initiates the study of TECCs and, in particular, analyzes a class of one-twist TECCs obtained from evaluation on rational points of a projective nonsingular elliptic curve over a finite field. For this class, the paper gives explicit parity-check matrices via Weil differentials, presents sufficient and necessary conditions of self-duality, determines the minimum distances, provides examples of MDS, AMDS, self-dual and MDS self-dual codes, and computes Schur-square dimensions to establish non-equivalence with elliptic curve codes and GRS codes [2509.03034].

## 1. Algebraic–geometric definition

Let \(q\) be a prime power, let \(E/\F_q\) be a projective nonsingular elliptic curve with function field \(F=\F_q(E)\), and write \(O\) for the unique point at infinity. Fix an integer \(k\ge 3\), choose \(n\) distinct rational points
\[
D=P_1+\cdots+P_n,\quad P_i\in E(\F_q)\setminus\{O\},
\]
choose an integer \(\ell\) satisfying
\[
0\;\le\;\ell\;\le \begin{cases} \frac{k-3}2,&k\text{ odd},\\[4pt] \frac{k}2,&k\text{ even}, \end{cases}
\]
and select a nonzero twist parameter \(\eta\in\F_q^*\) [2509.03034].

In affine coordinates \(E:y^2=f(x)\), or \(y^2+y=f(x)\) when \(p=2\), the construction uses the fact that \(\deg(O)=1\) and that \(L((k+1)O)\) has basis
\[
\{\,1,x,\dots,x^{\lfloor\frac{k+1}2\rfloor}\}\cup \{\,y,x\,y,\dots,x^{\lfloor\frac{k-1}2\rfloor}y\}.
\]
For odd \(k\), the one-twist defining set \(S_\ell\subset L((k+1)O)\) is
\[
S_\ell =\Bigl\{\sum_{i=0}^{\floor{\tfrac{k-1}2}}a_i\,x^i \;+\;\sum_{j=0}^{\floor{\tfrac{k-3}2}}b_j\,x^j\,y \;+\;b_\ell\,\eta\,x^{\tfrac{k+1}2}\Bigr\},
\]
with an analogous definition for even \(k\). The paper states that \(\dim_{\F_q}S_\ell=k\).

The associated TECC is the evaluation code
\[
\C\bigl(D,kO,\ell,\eta\bigr)\;=\;\bigl\{\,\bigl(f(P_1),\dots,f(P_n)\bigr)\;:\;f\in S_\ell\bigr\} \;\subset\;\F_q^n.
\]
This construction preserves the standard AG-code paradigm of evaluation on rational points while introducing a single twist term controlled by \(\eta\) and \(\ell\). A plausible implication is that the twist is designed to perturb the ambient function space without changing its dimension, thereby enabling code families with structural properties not shared by classical elliptic curve codes.

## 2. Basic parameters and minimum distance

The length of \(\C(D,kO,\ell,\eta)\) is \(n=\#\supp D\), and its dimension is \(k\) [2509.03034]. By Goppa’s bound,
\[
d\;\ge\;n-\deg\bigl((k+1)O\bigr)\;=\;n-(k+1).
\]

The paper strengthens this lower bound by proving that the minimum distance can assume only three values:
\[
d\in\{\,n-k-1,\;n-k,\;n-k+1\}.
\]
It then gives exact criteria for each case:
\[
\begin{array}{ll}
d=n-k-1 \iff \exists\{\!P_{i_1},\dots,P_{i_{\,k+1}\}\subset D\text{ with } \eta =\eta(\ell,P_{i_1},\dots,P_{i_{k+1}}), \\[6pt]
d=n-k \iff  \eta\neq\eta(\ell,P_{i_1},\dots,P_{i_{k+1}}) \;\forall\text{ size-}k+1\text{ subsets,} \;\;N(k,O,D)>0, \\[4pt]
d=n-k+1 \iff \eta\neq\eta(\ell,P_{i_1},\dots,P_{i_{k+1}}) \;\forall\text{ size-}k+1\text{ subsets,} \;\;N(k,O,D)=0,
\end{array}
\]
where \(N(k,O,D)\) is the usual subset-sum count in the group \(E(\F_q)\).

These conditions place the distance problem at the intersection of divisor theory and the group structure of \(E(\F_q)\). The appearance of \(N(k,O,D)\) indicates that the extremal behavior of the code depends not only on the degree of the divisor but also on additive configurations of the evaluation points on the elliptic curve. This suggests that TECCs inherit the arithmetic sensitivity of elliptic curve codes while adding twist-dependent obstructions through the parameter \(\eta\).

## 3. Dual codes and explicit parity-check matrices

A central structural result is the explicit determination of parity-check matrices through Weil differentials. Let
\[
t\;=\;\prod_{i=1}^n\!\bigl(x-\alpha_i\bigr),\qquad P_i=(\alpha_i,\beta_i),
\]
be a local uniformizer, and define the canonical differential
\[
\omega\;=\;\frac{dx}{t}.
\]
In the odd-characteristic model \(y^2=f(x)\), the paper shows that
\[
(\omega)=-D+nO+(y).
\]

Using Serre duality, or equivalently the standard AG-code duality formula, one obtains
\[
\C(D,kO,\ell,\eta)^\perp =\gamma\;\star\; \C\bigl(D,\,(y)+(n-k)O\bigr), \quad \gamma_i=\res_{P_i}(\omega).
\]
The relevant Riemann–Roch space admits the basis
\[
\{\,x^i/y:i=0,\dots,\floor{\tfrac{n-k}2}\}\cup \{\,x^j:j=0,\dots,\floor{\tfrac{n-k-3}2}\}.
\]
Accordingly, the rows of a parity-check matrix can be written in block form as
\[
H \;=\; \begin{pmatrix}
\gamma_i\,x_i^0/y_i& \gamma_i\,x_i^1/y_i& \cdots& \gamma_i\,x_i^{\floor{\tfrac{n-k}2}}/y_i \\
\gamma_i\,x_i^0& \gamma_i\,x_i^1& \cdots& \gamma_i\,x_i^{\floor{\tfrac{n-k-3}2}}
\end{pmatrix}_{i=1,\dots,n},
\]
and, after substituting \(y_i=\beta_i\), the paper gives the fully explicit form
\[
H\;=\; \begin{pmatrix}
\gamma_1\beta_1&\gamma_2\beta_2&\cdots&\gamma_n\beta_n \\
\gamma_1\alpha_1\beta_1&\gamma_2\alpha_2\beta_2&\cdots& \gamma_n\alpha_n\beta_n \\
\vdots&&&\vdots\\
\gamma_1\alpha_1^{\floor{(n-k)/2}}\beta_1& \cdots& \gamma_n\alpha_n^{\floor{(n-k)/2}}\beta_n \\
\gamma_1&\gamma_2&\cdots&\gamma_n \\
\gamma_1\alpha_1&\cdots&\gamma_n\alpha_n \\
\vdots&\vdots&\vdots&\vdots\\
\gamma_1\alpha_1^{\floor{(n-k-3)/2}}& \cdots& \gamma_n\alpha_n^{\floor{(n-k-3)/2}}
\end{pmatrix}.
\]

The significance of this description is methodological as well as structural. In classical AG coding theory, dual codes are often described abstractly through divisors and differentials; here, the paper pushes the description to an explicit matrix level. This makes the duality theory directly usable for parameter verification and for the study of self-duality.

## 4. Euclidean self-duality

The paper introduces a scaling vector \(\mathbf v=(v_1,\dots,v_n)\in(\F_q^*)^n\) and defines the scaled code \(\C(D,kO,\ell,\eta,\mathbf v)\) by
\[
(f(P_1),\dots,f(P_n))\;\longmapsto\;(v_1\,f(P_1),\dots,v_n\,f(P_n)).
\]
For one-twist TECCs with \(n=2k\), Theorem 9 gives a complete characterization of Euclidean self-duality [2509.03034].

The code is Euclidean self-dual if and only if
\[
\Bigl(\sum_{i=1}^n\gamma_i\beta_i\alpha_i^{k+1}\Bigr)\,\eta \;+\; 2\,\sum_{i=1}^n\gamma_i\alpha_i^{k-1} \;=\;0,
\]
and simultaneously there exists \(\lambda\in\F_q^*\) such that
\[
v_i^2 \;=\;\lambda\,\gamma_i\,\beta_i, \quad i=1,\dots,n.
\]
These two equations are necessary and sufficient for
\[
\C(D,kO,\ell,\eta,\mathbf v)=\C(D,kO,\ell,\eta,\mathbf v)^\perp.
\]

The first condition couples the twist parameter \(\eta\) to explicit residue-weighted sums of curve coordinates, while the second imposes a quadratic compatibility on the scaling vector. This identifies self-duality not as a generic feature of the construction, but as a constrained arithmetic phenomenon depending on the evaluation set, the residues of the chosen differential, and the twist. A plausible implication is that TECCs provide a controlled framework in which self-dual AG codes can be produced by solving explicit algebraic conditions rather than relying solely on existential arguments.

## 5. Examples and extremal behavior

The paper provides concrete examples of MDS, AMDS, self-dual and MDS self-dual TECCs [2509.03034]. These examples instantiate the general theory and show that the one-twist construction can realize several standard optimality notions.

| Type | Parameters and field | Defining data |
|---|---|---|
| MDS self-dual | \([6,3,4]\) over \(\F_4\) | \(\F_4=\F_2(\alpha)\), \(E:y^2+y=x^3\), \(D=\sum_{i=1}^6P_i\), \(k=3\), \(\ell=0\), and \(\mathbf v\) chosen so that \(v_i^2=\lambda\,\gamma_i\beta_i\) |
| AMDS near-MDS | \([8,3,4]\) over \(\F_5\) | \(E:y^2=x^3+x+1\), \(n=8\), \(k=3\), twist \(\eta\in\{1,4\}\subset\F_5\) |

In the first example, the resulting \([6,3,4]\) code is both MDS and self-dual. In the second, one checks that \(d=4=n-k\), so the code is almost-MDS but not MDS. The paper also states that numerous further MDS, AMDS, self-dual and MDS self-dual examples are given in Sections 5–6 by carrying out the group-sum criteria \(N(k,O,D)\) and related conditions.

These examples are significant because they show that the twist mechanism does not merely perturb known AG constructions but can be tuned to preserve or create extremal metric and duality properties. This suggests that the family is broad enough to interpolate between classical elliptic-curve behavior and genuinely new code structures.

## 6. Schur square and non-equivalence

Let \(\C^{\star 2}=\C\star\C\) denote the component-wise Schur square. Theorem 14 shows that for
\[
4\;\le\;k\;\le\;\frac{n-4}2,
\]
the one-twist TECCs satisfy
\[
\dim\bigl(\C(D,kO,\ell,\eta)^{\star2}\bigr) \;\ge\;2k+1.
\]
By contrast, any \(k\)-dimensional GRS code or elliptic curve code satisfies
\[
\dim\bigl(\C^{\star2}\bigr)\le 2k
\]
[2509.03034].

From this, the paper concludes that for \(4\le k\le(n-4)/2\), the TECCs cannot be monomially equivalent to any GRS code or classical elliptic curve code; likewise, their duals are excluded when \((n+4)/2\le k\le n-4\). This is one of the strongest structural distinctions established in the paper. Rather than relying on informal intuition that a twist should generate a new family, the argument uses a Schur-square invariant to separate TECCs from two of the most prominent evaluation-code classes.

This addresses a common potential misconception: although TECCs are built from elliptic curves and evaluation maps, they are not merely reparameterized elliptic curve codes. In the parameter range covered by the theorem, the Schur-square dimension provides a formal obstruction to monomial equivalence. A plausible implication is that the twist changes the multiplicative profile of the code in a way that is invisible to basic parameters such as length and dimension but detectable through higher-order invariants.

## 7. Position within algebraic–geometric coding theory

The one-twist TECC construction sits within AG coding theory but is explicitly framed as an extension motivated by twisted generalized Reed-Solomon codes [2509.03034]. Its defining features are the use of a modified \(k\)-dimensional subspace \(S_\ell\subset L((k+1)O)\), the explicit dual description via Weil differentials, the exact minimum-distance trichotomy \(d\in\{n-k-1,n-k,n-k+1\}\), the necessary and sufficient self-duality conditions for the case \(n=2k\), and the Schur-square criterion establishing non-equivalence with GRS and classical elliptic curve codes.

Taken together, these results delineate TECCs as a distinct family of evaluation codes on elliptic curves. The construction retains the arithmetic and geometric apparatus of AG codes—divisors, Riemann–Roch spaces, rational points, residues, and duality via differentials—while altering the defining space by a single twist term. This suggests that TECCs provide a systematic way to enlarge the design space of elliptic-curve-based codes without abandoning the explicit algebraic structure that makes AG codes analyzable.

Source: https://www.emergentmind.com/topics/twisted-elliptic-curve-codes-teccs