Weierstrass semigroups at several points, total inflection points on - - PowerPoint PPT Presentation

weierstrass semigroups at several points total inflection
SMART_READER_LITE
LIVE PREVIEW

Weierstrass semigroups at several points, total inflection points on - - PowerPoint PPT Presentation

Weierstrass semigroups at several points, total inflection points on curves and coding theory C cero Carvalho Faculdade de Matem atica Universidade Federal de Uberl andia Algebraic curves over finite fields RICAM - November 1115,


slide-1
SLIDE 1

Weierstrass semigroups at several points, total inflection points on curves and coding theory

C´ ıcero Carvalho

Faculdade de Matem´ atica Universidade Federal de Uberlˆ andia

Algebraic curves over finite fields RICAM - November 11–15, 2013

Partially supported by CNPq, FAPEMIG and CAPES C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 1 / 12

slide-2
SLIDE 2

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-3
SLIDE 3

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-4
SLIDE 4

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-5
SLIDE 5

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-6
SLIDE 6

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-7
SLIDE 7

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-8
SLIDE 8

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-9
SLIDE 9

Weierstrass semigroup at several points

Let X be a smooth complete irreducible curve of genus g ≥ 1 defined over a field F, assumed to be the full field of constants of F(X). Let P1, . . . , Pm be distinct rational points of X. Definition The Weierstrass semigroup at P1, . . . , Pm is defined as H = H(P1, . . . , Pm) := {(α1, . . . , αm) ∈ Nm

0 | ∃ f ∈ F(X) with

div∞(f ) = α1P1 + · · · + αmPm} Its systematic study was initiated by S. J. Kim and M. Homma in mid 90’s. They studied specially the case m = 2; investigated properties of H and its relationship with the theory of algebraic geometry (Goppa) codes. In a joint work with F. Torres we extended their results for any value of m, and also applied the results to obtain better lower bounds for the minimum distance of certain algebraic geometry codes. A similar application of these semigroups was recently done by Korchm´ aros and Nagy, which improved such bounds for certain codes previously studied by Matthews and Michel.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 2 / 12

slide-10
SLIDE 10

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-11
SLIDE 11

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-12
SLIDE 12

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-13
SLIDE 13

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-14
SLIDE 14

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-15
SLIDE 15

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-16
SLIDE 16

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-17
SLIDE 17

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-18
SLIDE 18

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-19
SLIDE 19

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-20
SLIDE 20

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-21
SLIDE 21

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-22
SLIDE 22

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-23
SLIDE 23

Weierstrass semigroup at several points

We will write n := (n1, . . . , nm) ∈ Nm

0 , ei ∈ Nm 0 for the m-tuple that has 1

in the i-th position and 0 in the others, L(n) := L(n1P1 + · · · + nmPm) and ℓ(n) := dim L(n).

  • Lemma. The following are equivalent:

(i) n ∈ H; (ii) ℓ(n) = ℓ(n − ei) + 1 for all i = 1, . . . , m; (iii) The linear system |n1P1 + · · · + nmPm| is base-point free. We call Nm

0 \ H the set of gaps of H, it is a finite set whose cardinality

may vary with P1, . . . , Pm. For example, if X is a hyperelliptic curve of genus 4, and m = 2 we get:

P2

✲ P1 q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q q

9 gaps if P1 and P2 are not W. points of X

P2

✲ P1 q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q

26 gaps if P1 and P2 are both W. points of X C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 3 / 12

slide-24
SLIDE 24

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

slide-25
SLIDE 25

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

slide-26
SLIDE 26

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

slide-27
SLIDE 27

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

slide-28
SLIDE 28

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-29
SLIDE 29

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-30
SLIDE 30

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-31
SLIDE 31

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-32
SLIDE 32

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-33
SLIDE 33

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-34
SLIDE 34

Weierstrass semigroup at several points

From now on we assume that #(F) ≥ m. Properties of H:

  • For all i = 1, . . . , m we get that a ∈ H(Pi) if and only if a.ei ∈ H.
  • Let (n1, . . . , nm), (p1, . . . , pm) ∈ H and set qi := max{ni, pi},

i = 1, . . . , m. Then (q1, . . . , qm) ∈ H. Define (n1, . . . , nm) (p1, . . . , pm) if ni ≤ pi ∀i = 1, . . . , m. Then is a partial order in Nm

0 .

Let i ∈ {1, . . . , m}, let ni ∈ N0 and let n = (n1, . . . , nm) be a minimal element (w.r.t. )

  • f the set {(p1, , . . . , pm) ∈ H | pi = ni}.

If ni > 0 and nj > 0 for some j ∈ {1, . . . , m}, j = i, then: (i) niei / ∈ H (hence ni / ∈ H(Pi)); (ii) n is a minimal element

  • f

the set {(p1, . . . , pm) ∈ H | pj = nj}, so nj / ∈ H(Pj).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 4 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-35
SLIDE 35

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

slide-36
SLIDE 36

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

slide-37
SLIDE 37

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-38
SLIDE 38

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-39
SLIDE 39

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛ q q

∇2(n)

slide-40
SLIDE 40

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q q

slide-41
SLIDE 41

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q q q q q q

∇2(n) ∇1(n)

slide-42
SLIDE 42

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-43
SLIDE 43

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-44
SLIDE 44

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-45
SLIDE 45

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q

slide-46
SLIDE 46

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-47
SLIDE 47

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-48
SLIDE 48

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-49
SLIDE 49

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-50
SLIDE 50

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-51
SLIDE 51

Weierstrass semigroup and AG codes

Given n = (n1, . . . , nm) ∈ Nm

0 , define

∇i(n) := {(p1, . . . , pm) ∈ H | pi = ni and pj ≤ nj ∀j = i}

  • Lemma. Let n ∈ Nm

0 . The following are equivalent:

(i) n / ∈ H; (ii) ∇i(n) = ∅ for some i ∈ {1, . . . , m}. We say that n ∈ Nm

0 is a pure gap if ∇i(n) = ∅ for all i ∈ {1, . . . , m}.

Denote the set of pure gaps by G0. Lemma: Let n = (n1, . . . , nm) ∈ Nm

0 .

(i) If n ∈ G0 then ni / ∈ H(Pi) for all i = 1, . . . , m. (ii) If 1 +

i ni ≤ γ, where γ is the gonality of

X, then n ∈ G0.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 5 / 12

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ✻

P2 P1

✲ q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ ❛ q q q q ❛ ❛ ❛ q q q ❛ ❛ q ❛ q q q q q q q q ❛

slide-52
SLIDE 52

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-53
SLIDE 53

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-54
SLIDE 54

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-55
SLIDE 55

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-56
SLIDE 56

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-57
SLIDE 57

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-58
SLIDE 58

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-59
SLIDE 59

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-60
SLIDE 60

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-61
SLIDE 61

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-62
SLIDE 62

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-63
SLIDE 63

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-64
SLIDE 64

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-65
SLIDE 65

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12

slide-66
SLIDE 66

Weierstrass semigroup and AG codes

Assume that F is a finite field, let D := Q1 + · · · + Qn, where Q1, . . . , Qn are distinct rational points of X, all distinct from P1, . . . , Pm, and let G be a divisor with support on P1, . . . , Pm. Let CΩ(D, G) be the algebraic geometry code which is the image of the map ϕ : Ω(G − D) → Fn defined by ϕ(η) = (resQ1(η), . . . , resQn(η)). We know that CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2). Building on previous work by Homma, Kim and Matthews, F. Torres and myself proved the following results.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

and let G = m

i=1(ni + pi − 1)Pi. Then CΩ(D, G) is an [n, k, d]-code,

with d ≥ deg(G) − (2g − 2) + m.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G = m

i=1(ni + pi − 1)Pi,

then CΩ(D, G) is an [n, k, d]-code, with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 6 / 12 Weierstrass semigroup of two rational points in Y 8 + Y = X 9 over F64.

slide-67
SLIDE 67

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-68
SLIDE 68

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-69
SLIDE 69

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-70
SLIDE 70

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-71
SLIDE 71

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-72
SLIDE 72

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-73
SLIDE 73

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-74
SLIDE 74

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-75
SLIDE 75

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-76
SLIDE 76

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-77
SLIDE 77

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-78
SLIDE 78

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-79
SLIDE 79

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-80
SLIDE 80

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-81
SLIDE 81

Weierstrass semigroup and AG codes

A similar improvement can be made to bounds for the generalized Hamming distance of AG codes. Let r be a positive integer, and C ⊂ Fm a linear code. Let U be a subcode

  • f C, the support of U is defined as

supp(U) := {i | ci = 0 for some (c1, . . . , cm) ∈ U}. The r-th generalized Hamming distance of C is defined as dr(C) = min{#(supp(U)) | U is a subcode of C, dim(U) = r}. Let s be a positive integer and set γs := {min(deg(A)) | A a divisor with dim L(A) = s}. The sequence γ1, γ2, ... is the gonality sequence of X; γ1 = 0 and γ2 is the gonality of X. This concept was introduced by Yang, Kummar and Stichtenoth in a paper where they proved the following result.

  • Theorem. The r-th generalized Hamming distance of an AG code of

length n defined over X satisfies i) dr(CL(D, G) ≥ n − deg(G) + γr; ii) dr(CΩ(D, G) ≥ deg(G) − (2g − 2) + γr.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 7 / 12

slide-82
SLIDE 82

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-83
SLIDE 83

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-84
SLIDE 84

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-85
SLIDE 85

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-86
SLIDE 86

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-87
SLIDE 87

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-88
SLIDE 88

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-89
SLIDE 89

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-90
SLIDE 90

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-91
SLIDE 91

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-92
SLIDE 92

Weierstrass semigroup and AG codes

Using the concept of pure gaps we were able to prove the following result.

  • Theorem. Assume that (n1, . . . , nm) and (p1, . . . , pm) are pure gaps of H,

with ni ≤ pi for all i = 1, . . . , m, and that (q1, . . . , qm) is also a pure gap whenever ni ≤ qi ≤ pi, for all i = 1, . . . , m. Let G := m

i=1 piPi. Then:

i) dr(CL(D, G) ≥ n − deg(G) + m

i=1(pi − ni) + m + γr;

ii) dr(CΩ(D, G) ≥ deg(G)−(2g −2)+γr+m

i=1(pi−ni)+m−(m

i=1(pi −ni)+m).

The last item can lead to an improvement because γr < γr+1 for all positive integers r. Thus the existence of pure gaps in H may lead to an improvement of the bounds for the generalized Hamming distances of AG codes. Now we will show that the existence of total inflection points in plane curves determine the existence of pure gaps in certain Weierstrass semigroups.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 8 / 12

slide-93
SLIDE 93

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-94
SLIDE 94

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-95
SLIDE 95

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-96
SLIDE 96

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-97
SLIDE 97

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-98
SLIDE 98

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-99
SLIDE 99

Total inflection points and pure gaps

Assume that X is a smooth plane, projective curve, of degree r > 4. We say that P ∈ X is a total inflection point if the tangent line at P intersects X only at P. In a work with T. Kato, we proved the following.

  • Theorem. Let P1, P2 and P3 be rational, total inflection points of X

which do not lie in a line. Then ((r − 4)r, 1, 1), (1, (r − 4)r, 1) and (1, 1, (r − 4)r) are pure gaps of H(P1, P2, P3).

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

j=1 sj, for all

i = 1, . . . , m, and m

j=1 sj ≤ r − 2 − m.

  • Theorem. Let P, P1, . . . , Pm ∈ X be rational points, with P a total

inflection point. Let 0 ≤ i < r − 3 and α1, . . . , αm be positive integers such that m

j=1 αj ≤ r − i − 3. Then (ir + 1, α1, . . . , αm) is a pure gap of

H(P, P1, . . . , Pm).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 9 / 12

slide-100
SLIDE 100

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-101
SLIDE 101

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-102
SLIDE 102

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-103
SLIDE 103

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-104
SLIDE 104

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-105
SLIDE 105

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-106
SLIDE 106

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-107
SLIDE 107

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-108
SLIDE 108

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ deg(G) − (2g − 2) + m + n

i=1(pi − ni)

d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-109
SLIDE 109

Total inflection points and pure gaps

  • Theorem. Let P1, . . . , Pm be total inflection points on X. Then

(s1r + α1, . . . , smr + αm) is a pure gap of H(P1, . . . , Pm), whenever si, αi are integers such that si ≥ 0, 1 ≤ αi ≤ r − 1 − i − m

i=1 si, for all

i = 1, . . . , m, and m

i=1 si ≤ r − 2 − m.

  • Application. Take X ⊂ P2(K) the Hermitian curve of degree q + 1 defined
  • ver F = GF(q2).

Let s and m be positive integers such that s + m ≤ q − 1; let P1, . . . , Pm be distinct rational points of X. Take s1 = s, s2 = · · · = sm = 0, from the above theorem we get that (sr + α1, α2, . . . , αm) is a pure gap at H(P1, . . . , Pm) whenever 1 ≤ αi ≤ q − i − s (i = 1, . . . , m) (i.e. only have pure gaps “between” (sr + 1, 1, . . . , 1) and (sr + q − 1 − s, q − 2 − s, . . . , q − m − s)). Let G = (2sr + q − 1 − s)P1 + m

i=2(q − i − s)Pi and let D be the sum of

the other q3 + 1 − m rational points of X. From the work together with F. Torres we know that CΩ(D, G) is an [q3 + 1 − m, k, d] code with d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2).

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 10 / 12

slide-110
SLIDE 110

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-111
SLIDE 111

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-112
SLIDE 112

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-113
SLIDE 113

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-114
SLIDE 114

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-115
SLIDE 115

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-116
SLIDE 116

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-117
SLIDE 117

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-118
SLIDE 118

Total inflection points and pure gaps

So d ≥ 2s(q + 1) + m(2q − 2s − 1) − m2 − (q + 1)(q − 2) and if we take s ≥ (q − 1)/2 then deg(G) > 2g − 2 and k = g + q3 − 2s(q + 1)− m(q − s) + m(m − 1)/2. We compared CΩ(D, G) with codes on the Hermitian curve supported on

  • ne point and having the same dimension k, finding many situations where

CΩ(D, G) has better parameters. For example, asssume that q is odd and q ≥ 5, take m = s = (q − 1)/2. Then CΩ(D, G) is an [q3 + 1 − (q − 1)/2, k, d]-code with k = q3 − (5q + 13)(q − 1)/8 e d ≥ q2/4 + q + 3/4. Taking F = (q3 − (q2/8 + 3q/2 − 5/8))P, where P is a rational point of X and E is the sum of the other rational points, we get that CΩ(F, E) is an [q3, k, d′] code, where d′ = q2/8 + 3q/2 − 5/8 (from works by Stichtenoth, Yang and Kummar) so that d − d′ ≥ (q(q − 4) + 11)/8.

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 11 / 12

slide-119
SLIDE 119

T H A N K Y O U!

C´ ıcero Carvalho (UFU) Inflection points on curves and coding theory 12 / 12