Characterisation of Parallel Independence in AGREE-Rewriting Michael - - PowerPoint PPT Presentation

β–Ά
characterisation of parallel independence in agree
SMART_READER_LITE
LIVE PREVIEW

Characterisation of Parallel Independence in AGREE-Rewriting Michael - - PowerPoint PPT Presentation

Characterisation of Parallel Independence in AGREE-Rewriting Michael Lwe (FHDW Hannover) ICGT 2018, Toulouse June 26, 2018 Contents Partial arrow classifier AGREE-rewriting Gluing construction Residual Parallel independence


slide-1
SLIDE 1

Characterisation of Parallel Independence in AGREE-Rewriting

Michael LΓΆwe (FHDW Hannover)

ICGT 2018, Toulouse June 26, 2018

slide-2
SLIDE 2

ICGT 2018

Contents

Partial arrow classifier AGREE-rewriting Gluing construction Residual Parallel independence Characterisation

2

slide-3
SLIDE 3

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

3

slide-4
SLIDE 4

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

4

slide-5
SLIDE 5

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

4

slide-6
SLIDE 6

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

5

O

slide-7
SLIDE 7

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

5

O

πœƒO

O●

slide-8
SLIDE 8

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

5

O

πœƒO

O● D X

i f

slide-9
SLIDE 9

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

5

O

πœƒO

O● D X

i f (i, f )●

Pullback

slide-10
SLIDE 10

ICGT 2018

Partial Arrow Classifier

A category has partial arrow classifiers, if the following

  • bject-indexed family of morphisms exists:

For every object O, there is a monomorphism πœƒO: O β†’ O● which satisfies the following universal property: For every pair of morphisms (i : D β†’ X, f : D β†’ O) with monic i, there is a unique morphism (i, f )● : X β†’ O● such that the pair (i, f ) is pullback of the pair (πœƒO, (i, f )● ).

5

O

πœƒO

O● D X

i f (i, f )●

Pullback Partial arror Partial arrow

slide-11
SLIDE 11

ICGT 2018

Partial Arrow Classifier

6

O

πœƒO

O● X D

i f (i, f )●

Pullback

slide-12
SLIDE 12

ICGT 2018

πœƒO

O

id

Partial Arrow Classifier

6

O O● X

(i, f )●

Pullback

D

i f f

slide-13
SLIDE 13

ICGT 2018

πœƒO

O

id

7

O O● X

(i, f )●

Pullback

D

i f f

Partial Arrow Classifier

slide-14
SLIDE 14

ICGT 2018

πœƒO

O

id

7

O O● X

(i, f )●

Pullback

D

i f f id

X

i

= = [i, f ] = [πœƒO, id] β—¦ [id, (i, f )●] = = [πœƒO, id] β—¦ πœ… ((i, f )●) Partial Arrow Classifier

slide-15
SLIDE 15

ICGT 2018 8

πœƒO

O

id

O O● X

(i, f )●

Pullback

D

i f f id

X

i

= = [i, f ] = [πœƒO, id] β—¦ [id, (i, f )●] = = [πœƒO, id] β—¦ πœ… ((i, f )●) Partial Arrow Classifier

slide-16
SLIDE 16

ICGT 2018 8

πœƒO

O

id

O O● X

(i, f )●

Pullback

D

i f f id

X

i

= = [i, f ] = [πœƒO, id] β—¦ [id, (i, f )●] = = [πœƒO, id] β—¦ πœ… ((i, f )●) Partial Arrow Classifier

A category has partial arrow classifiers, if for every object O there is object O● and partial morphism 𝜁O: O● β†’ O such that for every object X and partial morphism (p : X β†’ O) there is unique total morphism p● : X β†’ O● with 𝜁O β—¦ πœ… (p●) = p, where functor πœ… is given by: πœ…O : O ⟼ O and πœ…M : m ⟼ [id, m]. The embedding from category C (with total arrows) into the category of partial arrows over C is a free construction!

slide-17
SLIDE 17

ICGT 2018 8

πœƒO

O

id

O O● X

(i, f )●

Pullback

D

i f f id

X

i

= = [i, f ] = [πœƒO, id] β—¦ [id, (i, f )●] = = [πœƒO, id] β—¦ πœ… ((i, f )●) Partial Arrow Classifier

A category has partial arrow classifiers, if for every object O there is object O● and partial morphism 𝜁O: O● β†’ O such that for every object X and partial morphism (p : X β†’ O) there is unique total morphism p● : X β†’ O● with 𝜁O β—¦ πœ… (p●) = p, where functor πœ… is given by: πœ…O : O ⟼ O and πœ…M : m ⟼ [id, m]. The embedding from category C (with total arrows) into the category of partial arrows over C is a free construction!

Pushouts are hereditary Pushouts preserve monomorphism Pushouts along monomorphisms are pullbacks Category has epi-mono-factorisation Pullbacks are preserved by embedding ……

slide-18
SLIDE 18

ICGT 2018

Partial Arrow Classifier: Set

9

1 2 3 4

slide-19
SLIDE 19

ICGT 2018

Partial Arrow Classifier: Set

9

✘

1 2 3 4 1 2 3 4

πœƒ

slide-20
SLIDE 20

ICGT 2018

Partial Arrow Classifier: Set

9

✘

1 2 3 4 1 2 3 4

πœƒ

5 6 31 42 32 41 43 7 31 42 32 41 43

f i

slide-21
SLIDE 21

ICGT 2018

Partial Arrow Classifier: Set

9

✘

1 2 3 4 1 2 3 4

πœƒ

5 6 31 42 32 41 43 7 31 42 32 41 43

f i (i, f )●

✘ ✘ ✘

Pullback

slide-22
SLIDE 22

ICGT 2018

Partial Arrow Classifier: Graph

10

1 2

slide-23
SLIDE 23

ICGT 2018

Partial Arrow Classifier: Graph

10

1 2

πœƒ

1 2

✘

slide-24
SLIDE 24

ICGT 2018

3 4

Partial Arrow Classifier: Graph

10

1 2

πœƒ

1 2

✘

f i

1 21 21 1 21 21

slide-25
SLIDE 25

ICGT 2018

✘ ✘

Partial Arrow Classifier: Graph

10

1 2

πœƒ

1 2

(i, f )●

✘

f i

1 21 21 1 21 21

Pullback

slide-26
SLIDE 26

ICGT 2018

Partial Arrow Classifier: OO-Model

11

A E C

slide-27
SLIDE 27

ICGT 2018

Partial Arrow Classifier: OO-Model

11

A E C A E C ✘

πœƒ

slide-28
SLIDE 28

ICGT 2018

Partial Arrow Classifier: OO-Model

11

A E C A E C ✘

πœƒ

A1 E C A2

f

A1 E C A2

i

S1 S2

slide-29
SLIDE 29

ICGT 2018

Partial Arrow Classifier: OO-Model

11

A E C A E C ✘

πœƒ (i, f )●

Pullback

A1 E C A2

f

A1 E C A2

i

✘ ✘

slide-30
SLIDE 30

ICGT 2018

Partial Arrow Classifier: OO-Model

11

A E C A E C ✘

πœƒ (i, f )●

Pullback

A1 E C A2

f

A1 E C A2

i

✘ ✘

slide-31
SLIDE 31

ICGT 2018

β€šInversionβ€˜ of Matches

12

A E C A E C ✘

πœƒ (m, id)●

Pullback

A E C

id

A E C ✘

m

✘ ✘ ✘

slide-32
SLIDE 32

ICGT 2018

β€šInversionβ€˜ of Matches

12

A E C A E C ✘

πœƒ

Pullback

A E C A E C ✘ ✘ ✘ ✘

m m●

slide-33
SLIDE 33

ICGT 2018

β€šInversionβ€˜ of Matches

13

A E C A E C ✘

πœƒ

C A E

m’

slide-34
SLIDE 34

ICGT 2018

β€šInversionβ€˜ of Matches

13

A E C A E C ✘

πœƒ

Pullback

m’ m’●

✘ ✘ ✘ ✘ C A E

slide-35
SLIDE 35

ICGT 2018

AGREE-Rewriting

14

slide-36
SLIDE 36

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule:

slide-37
SLIDE 37

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule: L● πœƒL (t, l)● (PB)

slide-38
SLIDE 38

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule: L● πœƒL (t, l)● (PB) G Base Match: m m● Inverse Match:

slide-39
SLIDE 39

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule: L● πœƒL (t, l)● (PB) G Base Match: m m● Inverse Match: n’ D g n

(PB)

(PB)

slide-40
SLIDE 40

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule: L● πœƒL (t, l)● (PB) G Base Match: m m● Inverse Match: n’ D g n p’ h H (PO)

(PB)

(PB)

slide-41
SLIDE 41

ICGT 2018

AGREE-Rewriting

14

T t L K R l r Rule: L● πœƒL (t, l)● R’ t’ r’ Context Rule: (PB) (PO) G Base Match: m m● Inverse Match: n’ D g n p’ h H (PO)

(PB) (FPC)

Trace: p h H (PB)

slide-42
SLIDE 42

ICGT 2018

AGREE: Practical Example

15

Extract abstract type (Version 1):

slide-43
SLIDE 43

ICGT 2018

AGREE: Practical Example

16

Extract abstract type (Version 2):

slide-44
SLIDE 44

ICGT 2018

AGREE: Local Copies

17

L●

x

L K T

x

πœƒ l t (t, l)● r n

slide-45
SLIDE 45

ICGT 2018

AGREE: Local Copies

17

L●

x

L K T

x

πœƒ l t (t, l)●

G

m r m●

x x

n

slide-46
SLIDE 46

ICGT 2018

AGREE: Local Copies

17

L●

x

L K T

x

πœƒ l t (t, l)●

G

m r m●

x x

n g

D

x x

n’

slide-47
SLIDE 47

ICGT 2018

AGREE: Global Copies

18

L●

x

L K T

x

πœƒ l t (t, l)● r n

x

slide-48
SLIDE 48

ICGT 2018

AGREE: Global Copies

18

L●

x

L K T

x

πœƒ l t (t, l)●

G

m r m●

x x

n g

D

x x

n’

x x x

slide-49
SLIDE 49

ICGT 2018

AGREE: Local Deletion

19

L●

x

L K T

x

πœƒ l t (t, l)● r n

slide-50
SLIDE 50

ICGT 2018

AGREE: Local Deletion

19

L●

x

L K T

x

πœƒ l t (t, l)●

G

m r m●

x x

n g

D

n’

x x

slide-51
SLIDE 51

ICGT 2018

AGREE: Global Deletion

20

L●

x

L K T

πœƒ l t (t, l)● r n

slide-52
SLIDE 52

ICGT 2018

AGREE: Global Deletion

20

L●

x

L K T

πœƒ l t (t, l)●

G

m r m●

x x

n g

D

n’

slide-53
SLIDE 53

ICGT 2018

AGREE: Local Addition

21

L●

x

L K T

πœƒ l t (t, l)● r

x

R

slide-54
SLIDE 54

ICGT 2018

AGREE: Local Addition

21

L●

x

L K T

πœƒ l t (t, l)●

G

m r m●

x x x

R

n g

D

n’

x x

D

x x

p’

slide-55
SLIDE 55

ICGT 2018

Gluing Construction

22

L K R l r G X m n

slide-56
SLIDE 56

ICGT 2018

Gluing Construction

22

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

(r, l) β—¦ (q, p) = (n, m) β—¦ (h, g)

slide-57
SLIDE 57

ICGT 2018

Gluing for DPO-Rewriting

23

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

slide-58
SLIDE 58

ICGT 2018

Gluing for DPO-Rewriting

23

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

Isomorphism Monomorphism Pushout Complement

slide-59
SLIDE 59

ICGT 2018

Gluing for SPO-Rewriting

24

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

Isomorphism Monomorphism

slide-60
SLIDE 60

ICGT 2018

Gluing for SqPO-Rewriting

25

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

Isomorphism

slide-61
SLIDE 61

ICGT 2018

Gluing for AGREE-Rewriting

26

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

slide-62
SLIDE 62

ICGT 2018

Gluing for AGREE-Rewriting

26

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

Special monomorphism Monomorphism Induced by base rule Induced by monic base match

slide-63
SLIDE 63

ICGT 2018

Gluing Construction

27

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

slide-64
SLIDE 64

ICGT 2018

Gluing Construction

27

L K R l r G X m n Z x u

(PB)

Y y p

(FPC)

D g v

(FPC)

H h q

(PO)

slide-65
SLIDE 65

ICGT 2018

Gluing Construction

28

L K R G X Z Y D H

slide-66
SLIDE 66

ICGT 2018

Gluing Construction

28

L K R G X Z Y D H K’ R’ K’’ Z’’ Y H’ D’’

slide-67
SLIDE 67

ICGT 2018

Gluing Construction

28

L K R G X Z Y D H K’ R’ Z’ D’ K’’ Z’’ Y H’ D’’

slide-68
SLIDE 68

ICGT 2018

Gluing Construction

28

L K R G X Z Y D H K’ R’ Z’ D’ K’’ Z’’ Y H’ D’’ Gluing diagrams compose and decompose like pushouts

slide-69
SLIDE 69

ICGT 2018

Parallel Independence

29

(l, r) (m’, n’) (m, n) (l’, r’) (g, h) (g’, h’) (p, q) (p’, q’)

slide-70
SLIDE 70

ICGT 2018

Parallel Independence

29

(l, r) (m’, n’) (m, n) (l’, r’) (g, h) (g’, h’) (p, q) (p’, q’)

(g’, h’) β—¦ (m, n) (g, h) β—¦ (m’, n’)

Residual? Residual?

slide-71
SLIDE 71

ICGT 2018

Parallel Independence

29

(l, r) (m’, n’) (m, n) (l’, r’) (g, h) (g’, h’) (p, q) (p’, q’)

(g’, h’) β—¦ (m, n) (g, h) β—¦ (m’, n’)

Residual? Residual?

slide-72
SLIDE 72

ICGT 2018

Parallel Independence

29

(l, r) (m’, n’) (m, n) (l’, r’) (g, h) (g’, h’) (p, q) (p’, q’)

(g’, h’) β—¦ (m, n) (g, h) β—¦ (m’, n’)

Residual? Residual?

slide-73
SLIDE 73

ICGT 2018

Residual

30

G H D g h

slide-74
SLIDE 74

ICGT 2018

Residual

30

G H D g h πœƒL L L● m m● mgh● mgh

slide-75
SLIDE 75

ICGT 2018

Residual

30

G H D g h πœƒL L L● m m● m’ g β—¦ m’ = m, h β—¦ m’ = mgh mgh● mgh

slide-76
SLIDE 76

ICGT 2018

Residual

30

G H D g h πœƒL L L● m m● m’ g β—¦ m’ = m, h β—¦ m’ = mgh m’● mgh● m● β—¦ g = m’●, mgh● β—¦ h = m’● mgh

slide-77
SLIDE 77

ICGT 2018

Residual

30

G H D g h πœƒL L L● m m● m’ m’● mgh● mgh

(m’, idL) pullback of (m, g) and (h, mgh)

slide-78
SLIDE 78

ICGT 2018

Residual

31

G L L●

x

πœƒL m m●

H D

mgh● mgh g h

slide-79
SLIDE 79

ICGT 2018

Residual

31

G L L●

x

πœƒL m m●

H D

mgh● mgh g h m’ g β—¦ m’ = m, h β—¦ m’ = mgh

slide-80
SLIDE 80

ICGT 2018

Residual

31

G L L●

x

πœƒL m m●

H D

mgh● mgh g h m’ m’●

x

g β—¦ m’ = m, h β—¦ m’ = mgh m● β—¦ g β‰  m’●, mgh● β—¦ h β‰  m’●

slide-81
SLIDE 81

ICGT 2018

Residual

32

G H D g h πœƒL L L● m m● m’ m’● mgh● mgh (m’, idL) pullback of (m, g) and (h, mgh)

slide-82
SLIDE 82

ICGT 2018

(m’, idL) pullback of (m, g) and h β—¦ m’ = mgh

Residual

32

G H D g h πœƒL L L● m m● m’ m’●

slide-83
SLIDE 83

ICGT 2018

Residual

33

G H D g h

slide-84
SLIDE 84

ICGT 2018

Residual

33

G H D g h m’ m’● (h β—¦ m’)● L πœƒL L● m m● K T l t (t, l)●

slide-85
SLIDE 85

ICGT 2018

Residual

33

G H D g h m’ m’● (h β—¦ m’)● L πœƒL L● m m● K T l t (t, l)● h’ w Y v X x y

(PB) (PB)

slide-86
SLIDE 86

ICGT 2018

Residual

33

G H D g h m’ m’● (h β—¦ m’)● L πœƒL L● m m● K T l t (t, l)● h’ w Y v X x y

(PB) (PB)

(FPC)

slide-87
SLIDE 87

ICGT 2018

Characterising Independence

34

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)●

slide-88
SLIDE 88

ICGT 2018

Characterising Independence

34

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)●

Match m1 for rule 1 has residual after applying rule 2 at m2, only if

  • 1. everything that m1 needs (locally copies, deletes, or

preserves) is neither copied nor deleted (neither locally nor globally) by rule 2 at match m2.

slide-89
SLIDE 89

ICGT 2018

Characterising Independence

35

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)●

slide-90
SLIDE 90

ICGT 2018

Characterising Independence

35

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 (𝜌2, 𝜌1)●

slide-91
SLIDE 91

ICGT 2018

Characterising Independence

35

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 (𝜌2, 𝜌1)●

(PB)

L1

idLI

slide-92
SLIDE 92

ICGT 2018

Characterising Independence

35

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 (𝜌2, 𝜌1)●

(PB)

L1

idLI

Match m1 for rule 1 has residual after applying rule 2 at m2, only if

  • 1. everything that m1 needs (locally copies, deletes, or

preserves) is neither copied nor deleted (neither locally nor globally) by rule 2 at match m2.

  • 2. everything that rule 1 adds is neither (globally)

copied nor deleted by rule 2 at match m2.

slide-93
SLIDE 93

ICGT 2018

Characterising Independence

35

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 (𝜌2, 𝜌1)●

(PB)

L1

idLI

Match m1 for rule 1 has residual after applying rule 2 at m2, only if

  • 1. everything that m1 needs (locally copies, deletes, or

preserves) is neither copied nor deleted (neither locally nor globally) by rule 2 at match m2.

  • 2. everything that rule 1 adds is neither (globally)

copied nor deleted by rule 2 at match m2.

add copy (globally) add copy (globally)

β‰ 

slide-94
SLIDE 94

ICGT 2018

Characterising Independence

36

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2

slide-95
SLIDE 95

ICGT 2018

Characterising Independence

36

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 K1 l1 r1 R1 D1 g1 n’1 m’2 m’’2 r1

slide-96
SLIDE 96

ICGT 2018

Characterising Independence

36

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 K1 l1 r1 R1 D1 g1 n’1 m’2 m’’2 (r1 β—¦ m’’2 , 𝜌1)● r1

slide-97
SLIDE 97

ICGT 2018

Characterising Independence

36

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 K1 l1 r1 R1 D1 g1 n’1 m’2 m’’2 (r1 β—¦ m’’2 , 𝜌1)● R1

idRI

(PB)

r1

slide-98
SLIDE 98

ICGT 2018

Characterising Independence

36

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 K1 l1 r1 R1 D1 g1 n’1 m’2 m’’2 (r1 β—¦ m’’2 , 𝜌1)● R1

idRI

(PB)

r1

Match m1 for rule 1 has residual after applying rule 2 at m2, if and only if

  • 1. everything that m1 needs (locally copies, deletes, or

preserves) is neither copied nor deleted (neither locally nor globally) by rule 2 at match m2.

  • 2. everything that rule 1 adds is neither (globally)

copied nor deleted by rule 2 at match m2.

slide-99
SLIDE 99

ICGT 2018

Characterising Independence

37

L1 G m1 m2 L2 K2 l2 L2● πœƒ2 T2 t2 (t2, l2)● L* 𝜌1 𝜌2 K1 l1 r1 R1 D1 g1 n’1 m’2 (𝜌2, 𝜌1)●

(PB)

L1

idLI

m’’2 (r1 β—¦ m’’2 , 𝜌1)● R1

idRI

(PB)

slide-100
SLIDE 100

ICGT 2018

Conclusion

AGREE-rewriting is instance of the Gluing Construction! There is a precise notion of residual! Gluing and mutual residuals provides Church-Rosser! Residuals can be characterized syntactically! β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€”β€” Are global effects useful?

38

slide-101
SLIDE 101

Thank you for your attention