Page 195 - DCAP305_PRINCIPLES_OF_SOFTWARE_ENGINEERING

Unit 9: Metrics

coupled different modules in a software system are. Different types of data bindings are possible Notes
[90]. A potential data binding is defined as a triplet (p,x,q), where p and q are modules and x is
a variable within the static scope of both p and q. This reflects the possibility that the modules
p and q may communicate with each other through the shared variable x. This binding does not
consider the internals of p and q to determine if the variable x is actually accessed in any of the
modules. This binding is based on data declaration. A used data binding is a potential binding
where both p and q use the variable x for reference or assignment. This is harder to compute
than potential data binding and requires more information about the internal logic of a module.
An actual data binding is a used data binding with the additional restriction that the module p
assigns a value to x and q references x. It is the hardest to compute, and it signifies the situation
where information may flow from the module p to module q through the shared variable x.
Computation of actual data binding requires detailed logic descriptions of modules p and 4.
All of these data bindings attempt to represent the strength of interconnections among modules.
The greater the number of bindings between two modules, the higher the interconnection between
these modules. For a particular type of binding, a matrix can be computed that contains the
number of bindings between different modules. This matrix can be used for further statistical
analysis to determine the interconnection strength of the system.
Always software metrics are based on code syntax, quantifying the complexity
or cohesion of an Implementation by performing calculations based on counting
code structures.
Self Assessment Questions

1. ............................... describe the logic structure of software modules.
(a) Flow graph notation (b) Cyclomatic complexity
(c) Graph matrices (d) Control flow graphs
2. ............................ can be characterized as the number of elements of a basis set of control
flow paths through the module.
(a) Flow graph notation (b) Cyclomatic complexity
(c) Graph matrices (d) Control flow graphs
3. ........................ are measures that capture the data interaction across portions of a software
system.
(a) Graph matrices (b) Potential data binding
(c) Data bindings (d) Control flow graphs
4. Rewriting a single multiway decision to cross a module boundary is not a clear violation
of structured design.
(a) True (b) False
5. The cyclomatic number in graph theory is defined as n + 1.
(a) n + 1 (b) n – 1
(c) (n + 1)2 (d) n(n + 1)

9.3 Cohesion Metric

Metrics can assist software developers and managers assess the quality of software and pinpoint
trouble areas in their systems. For instance, a metric may indicate that a class lacks cohesion. A
cohesive class is one in which all of the members are closely related, focused on a single task.
A class that lacks cohesion is poorly designed, and therefore is more likely to be error-prone.
If metrics are used to quickly and automatically find out which classes in software system lack
cohesion, the programming team can take steps to check those classes and improve them before

LOVELY PROFESSIONAL UNIVERSITY 189