Comp 587 Parker Li Bobby Kolski. Automated testing tools assist software engineers to gauge the quality of software by automating the mechanical aspects.

Comp 587

Parker Li

Bobby Kolski

Metrics for Measuring the Effectiveness of Software-Testing Tools

Automated testing tools assist software engineers to gauge the quality of software by automating the mechanical aspects of software-testing.

Automated testing tools vary in their:

• Underlying approach

• Quality

• Ease-of-use

How does a project manager choose the best suitable testing tool?

Introduction

• The history of software metrics began with counting the number of lines of code (LOC).

• It was assumed that more lines of code implied more complex programs, which in turn were more likely to have errors.

• However, software metrics have evolved well beyond the simple measures introduced in the 1960s.

Software Quality Metrics

• Since the first introduction of LOC, metrics for traditional or procedural source code have increased in number and complexity.

• Cyclomatic complexity

• e is the number of edges, n is the number of nodes and p is number of unconnected parts of G.

• Offers an estimate of the reliability, testability, and maintainability of a program, based on measuring the number of linearly independent paths through the program.

Procedural (Traditional) Software Metrics

Function Point is a measure of the size of computer applications and the projects that build them.

Procedural (Traditional) Software Metrics (Cont.)

• Weighted-methods-per-class (WMC) – the sum of the individual complexities of the methods within that class.

• Depth of inheritance tree (DIT) – defined as the maximum length from the node to the root of a class tree. Deeper a class is in the inheritance hierarchy, more complex it gets.

• Number of Children (NOC) – Immediate subclasses of class. Large NOC implies great amount of inheritance and resue.

• Coupling between object classes (CBO) – When a class inherits methods, instance variables, they are coupled. Greater number = complexity

Object-Oriented Software Metrics

• Number of key classes (NKC)

• Number of subsystems (NSUB)

• Class size (CS)

• Number of operations added by a subclass (NOA)

• Average method size

• Average number of instance variables

• Class hierarchy nesting level

Object-Oriented Software Metrics

• Does it offer critical support for planning tests and monitoring test progress?

• Does it have an analysis feature where it assesses the characteristics of the software quality?

• Does it offer guiding dynamic testing?

• How mature is the tool? Is it easy to implement for your system?

Which Software-Testing Tool to use?

• Human Interface Design (HID) – Tools with well-designed human interfaces enable easy, efficient, and accurate setting of tool configuration.

• Tool Management (TM) – the tool should ensure proper management of information.

• Ease of Use (EU) – the tool should be easy to use by both new and familiar users.

• User Control (UC) – the tool should allow extensive and detail testing coverage rather than one bulky one.

Metrics for Tools that support testing procedure software

3 Tools for Software Testing

•LDRA Testbed

•Parasoft Testbed

C++ Test

CodeWizard

Inspire++

•Telelogic Testbed

Logiscope

• Coverage Report (Dynamic Analysis)

Branch Coverage

Statement Coverage

• Metrics

Cyclomatic Complexity (Static Analysis)

• Quality Report

LRDA Testbed

LRDA Testbed – Branch/Decision coverage Part 1 (Part of dynamic testing)

LRDA Testbed – Code/branch/decision coverage Part 2 (Part of dynamic testing)

Parasoft Testbed

• C++ Test White-Box Black-Box Regression testing

• Code Wizard 170 industry-accepted C/C++ coding standards

•Inspire++ C/C++ runtime error checking

Parasoft Testbed (metrics)

Parasoft Testbed C++ Test Bug Detective (static analysis)

• Logiscope

• TestChecker – mentioned in the article

• RuleChecker – mentioned in the article

• Audit *

• Reviewer *

* included in the current version of the tool and shown for completeness but not covered further in the slides.

Telelogic Testbed

Telelogic Logiscope – Bought by IBM – The current version is Rational Telelogic Logiscope

• Human Interface Design (HID) Comparison

• Keyboard-to-mouse switches (KMS)

• Input fields per functions (IFPF)

• Average length of input fields (ALIF)

• Button recognition factor (BR)

Exercising the Software-Testing Tools (Comparison of the tools themselves)

Test case generation (TCG)

Automated test case generation (ATG)

Test case reuse functionality (TRF)

Special note: LDRA does not

automatically generate test

cases but does provide

user-friendly features such

as pull-down menus for created

test cases therefore it was

assigned an eight for its level of ATG.

Exercising the Software-Testing Tools (Comparison of the tools themselves)

LRDA – For Object Oriented tests there were issues with the tool, so no results were obtained.

High levels of knots indicate that the code is disjointed

LDRA appends the Motor Industry Software Reliability Association (MISRA) C Standard and the Federal Aviation Authority’s DO-178B standard.

Analysis of Results (The output of the tools)

• Software Metrics have come a long way since the LOC era.

• Software Metrics can be use not only to measure the quality of software, but also the effectiveness of the software testing tool.

• Choosing a software testing tool is not simple.

Importance of this article

Questions?