LECTURE 16: Class Cohesion Metrics

Ivan Marsic
Rutgers University 1
 Topics

• Structural Cohesion Metrics

• Internal Cohesion or Syntactic
Cohesion
• External Cohesion or Semantic
Cohesion

2
 Measuring Module Cohesion
• Cohesion or module “strength” refers to the
notion of a module level “togetherness”
viewed at the system abstraction level
• Internal Cohesion or Syntactic
Cohesion
– closely related to the way in which large programs are
modularized
– ADVANTAGE: cohesion computation can be automated
• External Cohesion or Semantic
Cohesion
– externally discernable concept that assesses whether the
abstraction represented by the module (class in object-oriented
approach) can be considered to be a “whole” semantically
– ADVANTAGE: more meaningful
3
 An Ordinal Cohesion Scale
6 - Functional cohesion high cohesion
module performs a single well-defined function
5 - Sequential cohesion
>1 function, but they occur in an order prescribed by the specification
4 - Communication cohesion
>1 function, but on the same data (not a single data structure or class)
3 - Procedural cohesion
multiple functions that are procedurally related
2 - Temporal cohesion
>1 function, but must occur within the same time span (e.g., initialization)
1 - Logical cohesion
module performs a series of similar functions, e.g., Java class java.lang.Math
0 - Coincidental cohesion low cohesion
PROBLEM: Depends on subjective human assessment 4
 Weak Cohesion
Indicates Poor Design
• Unrelated responsibilities/functions
imply that the module will have
unrelated reasons to change in the
future

• Because semantic cohesion is difficult

to automate, and automation is key,
most cohesion metrics focus on
syntactic cohesion
5
 Structural Class Cohesion

• SCC measures how well class

responsibilities are related
– Class responsibilities are expressed as its
operations/methods

• Cohesive interactions of class operations:

How operations can be related strongest cohesion
code 3 - Operations calling other operations (of this class)
based 2 - Operations sharing attributes
interface 1 - Operations having similar signatures (e.g., similar
based data types of parameters)
weakest cohesion
6
 Elements of a Software Class

Controller attributes
Code based cohesion metric:
a1: # numOfAttemps_ : long
a2: # maxNumOfAttempts_ : long a1 a2 To know if mi and mj are related, need to see their code
m1: + enterKey(k : Key)
m2: – denyMoreAttempts() m1 m2 Note: This is NOT strictly true, because good UML
interaction diagrams show which methods call other
methods methods, or which attributes are used by a method

Interface based cohesion metric:

To know if mi and mj are related, compare their signatures
a1

DeviceCtrl
a1: # devStatuses_ : Vector
m1: + activate(dev : string) : boolean m1 m2 m3
m2: + deactivate(dev :string) : boolean
m3: + getStatus(dev : string) : Object

Note:
A person can guess if a method is calling another method
or if a method is using an attribute,
but this process cannot be automated! 7
Interface-based Cohesion Metrics

• Advantages
– Can be calculated early in the design stage

• Disadvantages
– Relatively weak cohesion metric:
• Without source code, one does not know what exactly a method is
doing (e.g., it may be using class attributes, or calling other methods
on its class)
• Number of different classes with distinct method-attribute pairs is
generally larger than the number of classes with distinct method-
parameter-type, because the number of attributes in classes tends to be
larger than the number of distinct parameter types

8
 Desirable Properties
of Cohesion Metrics
• Monotonicity: adding cohesive interactions to the
module cannot decrease its cohesion
• if a cohesive interaction is added to the model, the modified model will exhibit
a cohesion value that is the same as or higher than the cohesion value of the
original model
• Ordering (“representation condition” of
measurement theory):
• Metric yields the same order as intuition
• Discriminative power (sensitivity): modifying
cohesive interactions should change the cohesion
– Discriminability is expected to increase as:
• 1) the number of distinct cohesion values increases and
• 2) the number of classes with repeated cohesion values decreases
• Normalization: allows for easy comparison of the
cohesion of different classes

9
 Example of 2 x 2 classes
List all possible cases for classes with two methods and two attributes.
We intuitively expect that cohesion increases from left to right:
attributes

a1 a2

m1 m2

methods

If we include operations calling other operations, then:

10
 Example of 2 x 2 classes
List all possible cases for classes with two methods and two attributes.
We intuitively expect that cohesion increases from left to right:
attributes

a1 a2

m1 m2

methods

If we include operations calling other operations, then:

11
 Cohesion Metrics
Running Example Classes
class C1 class C2 class C3
a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4

class C4 class C5 class C6

a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4

class C7 class C8 class C9

a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4

12
 class C1
Example Metrics (1)
class C2 class C3 class C4
a1 a4 a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4 m1 m4

Class Cohesion Metric Definition / Formula

(1)
Lack of Cohesion of Methods (LCOM1) LCOM1 = Number of pairs of methods that do not share attributes
(Chidamber & Kemerer, 1991)

P = Number of pairs of methods that do not share attributes P – Q, if P – Q  0

(2)
LCOM2
Q = Number of pairs of methods that share attributes LCOM2 =
(Chidamber & Kemerer, 1991)
0, otherwise

(3)
LCOM3 LCOM3 = Number of disjoint components in the graph that represents each method as a node and
(Li & Henry, 1993) the sharing of at least one attribute as an edge
C1, C4: C2: C3:

(4)
LCOM4 Similar to LCOM3 and additional edges are used to represent method invocations
(Hitz & Montazeri, 1995) C1: C2, C3: C4:

LCOM1(C1) = P = NP – Q = 6 – 1 = 5 LCOM2(C1) = P – Q = 5 – 1 = 4
M
LCOM1: LCOM1(C2) = 6 – 2 = 4 LCOM2(C2) = 4 – 2 = 2
# Method Pairs = NP = =
M!
LCOM1(C3) = 6 – 2 = 4 LCOM2: LCOM2(C3) = 4 – 2 = 2
2 2!  (M – 2)! LCOM1(C4) = 6 – 1 = 5 LCOM2(C4) = 5 – 1 = 4

LCOM3:
4 LCOM3(C1) = 3 LCOM4(C1) = 3
NP(Ci) = =6 LCOM3(C2) = 2
LCOM4: LCOM4(C2) = 2
2 LCOM3(C3) = 2 LCOM4(C3) = 2 13
LCOM3(C4) = 3 LCOM4(C4) = 1
 LCOM3 and LCOM4 for class C7
LCOM3 = Number of disjoint components in the graph that represents each method as a node and
the sharing of at least one attribute as an edge

class C7 Steps:
1. Draw four nodes (circles) for four methods.
a1 a4
2. Connect the first three circles because they are sharing attribute a1.

m1 m4
C7 & C7: LCOM3 creates the same graph for C7 and C7
m1 m2 m3 m4 --- there are two disjoint components in both cases

class C7 LCOM3(C7) = LCOM3(C7) = 2

a1 a4

LCOM4 = Similar to LCOM3 and additional edges are used to represent method invocations
m1 m4 Steps:
1. Draw four nodes (circles) for four methods.
2. Connect the first three circles because they are sharing attribute a1.
3. For C7 only: Connect the last two circles because m 3 invokes m4.

C7:
LCOM4 finds two disjoint components in case C7
LCOM4(C7) = 2
m1 m2 m3 m4

C7:
LCOM4 finds one disjoint component in case C7
LCOM4(C7) = 1
m1 m2 m3 m4 14
 Example Metrics (1)
class C7 class C8 class C9
a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4

Class Cohesion Metric Lack of Discrimination Anomaly (LDA) Cases

LDA1) When the number of method pairs that share common attributes is the same, regardless of how
many attributes they share, e.g., in C7 4 pairs share 1 attribute and in C8 4 pairs share 3 attributes each
(1)
Lack of Cohesion of Methods (LCOM1)
(Chidamber & Kemerer, 1991) LDA2) When the number of method pairs that share common attributes is the same, regardless of which
attributes are shared, e.g., in C7 4 pairs share same attribute and in C9 4 pairs share 4 different attributes
LDA1) and LDA2) same as for LCOM1
(2)
LCOM2
(Chidamber & Kemerer, 1991) LDA3) When P  Q, LCOM2 is zero, e.g., C7, C8, and C9

LDA1) same as for LCOM1

(3)
LCOM3
(Li & Henry, 1993) LDA4) When the number of disjoint components (have no cohesive interactions) is the same in the graphs
of compared classes, regardless of their cohesive interactions, e.g., inability to distinguish b/w C1 & C3

(4)
LCOM4 Same as for LCOM3
(Hitz & Montazeri, 1995)

class C1 LCOM1(C1) = P = NP – Q = 6 – 1 = 5 LCOM2(C1) = P – Q = 5 – 1 = 4

a1 a4 LCOM1(C3) = 6 – 2 = 4 LCOM2(C3) = 4 – 2 = 2
LCOM1: LCOM1(C7) = 6 – 3 = 3 LCOM2(C7) = 0 P<Q
LCOM1(C8) = 6 – 3 = 3 LCOM2: LCOM2(C8) = 0 P<Q
m1 m4 LCOM1(C9) = 6 – 3 = 3 LCOM2(C9) = 0 P<Q
class C3 LCOM3(C1) = 3 LCOM4(C1) = 3
LCOM3:
a1 a4 LCOM3(C3) = 2 LCOM4(C3) = 2
LCOM3(C7) = 2 LCOM4: LCOM4(C7) = 2
LCOM3(C8) = 2 LCOM4(C8) = 2 15
m1 m4
LCOM3(C9) = 1 LCOM4(C9) = 1
 LCOM3 and LCOM4 for class C7
LCOM3 = Number of disjoint components in the graph that represents each method as a node and
the sharing of at least one attribute as an edge

class C7 Steps:
1. Draw four nodes (circles) for four methods.
a1 a4
2. Connect the first three circles because they are sharing attribute a1.

m1 m4
C7 & C7: LCOM3 creates the same graph for C7 and C7
m1 m2 m3 m4 --- there are three disjoint components in both cases

class C7 LCOM3(C7) = LCOM3(C7) = 3

a1 a4

LCOM4 = Similar to LCOM3 and additional edges are used to represent method invocations
m1 m4 Steps:
1. Draw four nodes (circles) for four methods.
2. Connect the first three circles because they are sharing attribute a1.
3. For C7 only: Connect the last two circles because m 3 invokes m4.

C7:
LCOM4 finds three disjoint components in case C7
LCOM4(C7) = 3
m1 m2 m3 m4

C7:
LCOM4 finds one disjoint component in case C7
LCOM4(C7) = 1
m1 m2 m3 m4 16
 class C1
Example Metrics (2)
class C2 class C3 class C4
a1 a4 a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4 m1 m4

Class Cohesion Metric Definition / Formula

LCOM5 = (a – kℓ) / (ℓ – kℓ), where ℓ is the number of attributes, k is the number of methods, and a is the
(5)
LCOM5
(Henderson-Sellers, 1996) summation of the number of distinct attributes accessed by each method in a class

Coh = 1 – (1 – 1/k)LCOM5
Coh = a / kℓ, where a, k, and ℓ have the same definitions as above
(6)
Coh
k (1 – Coh)
(Briand et al., 1998) LCOM5 =
k–1

a(C1) = (2 + 1 + 1 + 1) = 5
a(C2) = (2 + 1 + 2 + 1) = 6
a(C3) = (2 + 2 + 1 + 1) = 6
a(C4) = (2 + 1 + 1 + 1) = 5

LCOM5(C1) = (5 – 44) / (4 – 44) = 11 / 12 Coh(C1) = 5 / 16

LCOM5: Coh:
LCOM5(C2) = 10 / 12 = 5 / 6 Coh(C2) = 6 / 16 = 3 / 8
LCOM5(C3) = 5 / 6 Coh(C3) = 3 / 8
LCOM5(C4) = 11 / 12 Coh(C4) = 5 / 16

17
 class C1
Example Metrics (2)
class C2 class C3 class C4
a1 a4 a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4 m1 m4

Class Cohesion Metric Lack of Discrimination Anomaly (LDA) Cases

(5)
LCOM5 LDA5) when classes have the same number of attributes accessed by methods, regardless of the
(Henderson-Sellers, 1996) distribution of these method-attribute associations, e.g., C2 and C3

Coh = 1 – (1 – 1/k)LCOM5
(6)
Coh Same as for LCOM5
k (1 – Coh)
(Briand et al., 1998) LCOM5 =
k–1

a(C1) = (2 + 1 + 1 + 1) = 5
a(C2) = (2 + 1 + 2 + 1) = 6
a(C3) = (2 + 2 + 1 + 1) = 6
a(C4) = (2 + 1 + 1 + 1) = 5

LCOM5(C1) = (5 – 44) / (4 – 44) = 11 / 12 Coh(C1) = 5 / 16

LCOM5: Coh:
LCOM5(C2) = 10 / 12 = 5 / 6 Coh(C2) = 6 / 16 = 3 / 8
LCOM5(C3) = 5 / 6 Coh(C3) = 3 / 8
LCOM5(C4) = 11 / 12 Coh(C4) = 5 / 16

18
 class C1
Example Metrics (3)
class C2 class C3 class C4
a1 a4 a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4 m1 m4

Class Cohesion Metric Definition / Formula

(7)
Tight Class Cohesion (TCC) TCC = Fraction of directly connected pairs of methods, where two methods are directly connected if
(Bieman & Kang, 1995) they are directly connected to an attribute. A method m is directly connected to an attribute when the
attribute appears within the method’s body or within the body of a method invoked by method m
directly or transitively C1: C3:
C2: C4:

(8)
Loose Class Cohesion (LCC) LCC = Fraction of directly or transitively connected pairs of methods, where two methods are
(Bieman & Kang, 1995) transitively connected if they are directly or indirectly connected to an attribute. A method m, directly
connected to an attribute j, is indirectly connected to an attribute i when there is a method directly or
transitively connected to both attributes i and j
C3:
In class C3: m1 and m3 transitively connected via m2 C1, C2: same as for TCC
C4:

(9)
Degree of Cohesion-Direct (DCD) DCD = Fraction of directly connected pairs of methods, where two methods are directly connected
(Badri, 2004) if they satisfy the condition mentioned above for TCC or if the two methods directly or transitively
invoke the same method C3:
C1, C2: same as for TCC
C4:
(10)
Degree of Cohesion-Indirect (DCI) DCI = Fraction of directly or transitively connected pairs of methods, where two methods are
(Badri, 2004) transitively connected if they satisfy the condition mentioned above for LCC or if the two methods
directly or transitively invoke the same method C3:
C1, C2: same as for TCC
C4:
NP – P LCOM1
TCC = Q* / NP = =1–
NP NP TCC(C1) = 1 / 6 LCC(C1) = 1/6 DCD(C1) = 1/6 DCI(C1) = 1/6
TCC(C2) = 2 / 6 LCC(C2) = 2/6 DCD(C2) = 2/6 DCI(C2) = 2/6
Q*(C4) = 3 TCC(C3) = 2 / 6 LCC(C3) = 3/6 DCD(C3) = 2/6 DCI(C3) = 3/6
NP(Ci) = 6 TCC: TCC(C4) = 3 / 6 LCC: LCC(C4) = 3/6 DCD: DCD(C4) = 4/6 DCI: DCI(C4) = 4/6 19
 class C1
Example Metrics (4)
class C2 class C3 class C4
a1 a4 a1 a4 a1 a4 a1 a4

m1 m4 m1 m4 m1 m4 m1 m4

Class Cohesion Metric Definition / Formula

CC = Ratio of the summation of the similarities between all pairs of methods to the total number
(11)
Class Cohesion (CC) of pairs of methods. The similarity between methods i and j is defined as:
(Bonja & Kidanmariam, 2006) | Ii  Ij |
Similarity(i, j) = where, Ii and Ij are the sets of attributes referenced by methods i and j
| Ii  Ij |

(12)
Class Cohesion Metric (SCOM) CC = Ratio of the summation of the similarities between all pairs of methods to the total number
(Fernandez & Pena, 2006) of pairs of methods. The similarity between methods i and j is defined as:
| Ii  Ij | | Ii  Ij |
Similarity(i, j) = . where, ℓ is the number of attributes
min(| Ii |, | Ij |) ℓ

0 if k = 0 or ℓ = 0
LSCC(C) = 1 if k = 1
(13)
Low-level design Similarity-based
ℓ
Class Cohesion (LSCC)
(Al Dallal & Briand, 2009)

i=1
x (x – 1)
i i

otherwise
ℓk (k – 1)
where ℓ is the number of attributes, k is the number of methods, and xi is the number of methods
that reference attribute i

CC(C1) = 1 / 2 SCOM(C1) = 2 / 4 = 1 / 2 LSCC(C1) = 2 / (4*4*3) = 2 / 48 = 1 / 24

CC(C2) = 1 SCOM(C2) = 2 / 4 + 2 / 4 = 1 LSCC(C2) = (2 + 2) / (4*4*3) = 1 / 12
CC(C3) = 1 SCOM(C3) = 2 / 4 + 2 / 4 = 1 LSCC(C3) = 1 / 12
CC: CC(C4) = 1 / 2 SCOM: SCOM(C4) = 2 / 4 = 1 / 2 LSCC: LSCC(C4) = 1 / 24 20
 Example Metrics (5)

Class Cohesion Metric Definition / Formula

CAMC = a/kℓ, where ℓ is the number of distinct parameter types, k is the number of methods,
(14)
Cohesion Among Methods and a is the summation of the number of distinct parameter types of each method in the class.
in a Class (CAMC) Note that this formula is applied on the model that does not include the “self” parameter type
(Counsell et al., 2006) used by all methods

 x (k – x )
2
NHD = 1 – , where k and ℓ are defined above for CAMC and xj
(15)
Normalized Hamming Distance (NHD) j j
(Counsell et al., 2006) ℓk (k – 1) j=1
is the number of methods that have a parameter of type j

(16)
Scaled Normalized Hamming SNHD = the closeness of the NHD metric to the maximum value of NHD compared to the
Distance (SNHD) minimum value
(Counsell et al., 2006)

21
 Cohesion Metrics
Performance Comparison

LCOM2

LCOM3

LCOM4

LCOM5
LCOM1

SCOM

LSCC
TCC

DCD
LCC
Coh

DCI

CC
class C1 5 4 3 3 11/12 5/16 1/6 1/6 1/6 1/6 1/2 1/2 1/24

class C2 4 2 2 2 5/6 3/8 2/6 2/6 1/3 1/3 1 1 1/12

class C3 4 2 2 2 5/6 3/8 2/6 2/6 1/3 1/2 1 1 1/12

class C4 5 4 3 1 11/12 5/16 1/6 3/6 2/3 2/3 1/2 1/2 1/24

class C5 5 4 3 1 1 1/4 1/6 3/6 4/6 4/6 1/2 1/2 1/24

class C6 3 0 1 1 2/3 1/2 1/2 4/6 3/6 4/6 - - -

class C7 3 0 2 2 13/12 3/16 1/2 2/6 2/6 2/6 - - -

or 3/6?
class C8 3 0 2 2 7/12 9/16 1/2 2/6 2/6 3/6 - - -
or 5/6?
class C9 3 0 1 1 2/3 1/2 1/2 3/6 3/6 5/6 - -
22
-