COMSATS Institute of Information Technology, Attock
1 Farhan Aadil
Farhan Aadil
Assistant Professor
COMSATS Institute of Information Technology
Lecture 13
Requirement Errors
COMSATS Institute of Information Technology, Attock
2 Farhan Aadil
Quiz 4
COMSATS Institute of Information Technology, Attock
3 Farhan Aadil
3
Today’s Topics
• Requirements errors
COMSATS Institute of Information Technology, Attock
4 Farhan Aadil
4
COMSATS Institute of Information Technology, Attock
5 Farhan Aadil
5
Requirements Error/Defect
COMSATS Institute of Information Technology, Attock
6 Farhan Aadil
6
Requirements Errors - 1
• Errors and omissions find their way in different requirements documents
COMSATS Institute of Information Technology, Attock
7 Farhan Aadil
7
Requirements Errors - 2
• It is difficult to detect requirements errors once they flow downstream
COMSATS Institute of Information Technology, Attock
8 Farhan Aadil
8
Types of Requirements Errors
• Errors of omission
• Errors of clarity and ambiguity
COMSATS Institute of Information Technology, Attock
9 Farhan Aadil
9
Errors of Omission
• Errors of omission are most common among requirements errors
COMSATS Institute of Information Technology, Attock
10 Farhan Aadil
10
Errors of Clarity and Ambiguity
• Second most common errors are those of clarity and ambiguity
• Primarily, because natural languages (like English) are used to state requirements, while such languages are themselves ambiguous
COMSATS Institute of Information Technology, Attock
11 Farhan Aadil
11
Performance Errors
• Performance, that is errors of speed and capacity, are also found in requirements
COMSATS Institute of Information Technology, Attock
12 Farhan Aadil
12
Negative Impact of Requirements Errors - 1
• The resulting software may not satisfy user’s real needs • Multiple interpretations of requirements may cause
COMSATS Institute of Information Technology, Attock
13 Farhan Aadil
13
Negative Impact of Requirements Errors - 2
• Negative impact on humans
– Unsatisfied customers and developers
COMSATS Institute of Information Technology, Attock
14 Farhan Aadil
14
Addressing Requirements Errors
COMSATS Institute of Information Technology, Attock
15 Farhan Aadil
15
Prevention vs. Removal
• For requirements errors, prevention is usually more effective than removal
• Joint application development (JAD), quality function
deployment (QFD), and prototyping are more effective in defect prevention
COMSATS Institute of Information Technology, Attock
16 Farhan Aadil
16
Defect Prevention - 1
• Don’t let defects/errors become part of the requirements document or requirements model in the first place
• How is it possible?
COMSATS Institute of Information Technology, Attock
17 Farhan Aadil
17
Defect Prevention - 2
• Training in different requirements engineering activities (elicitation, analysis and negotiation, specification, and validation) is also very important for defect prevention
COMSATS Institute of Information Technology, Attock
18 Farhan Aadil
18
Defect Prevention - 3
COMSATS Institute of Information Technology, Attock
19 Farhan Aadil
19
Defect Prevention - 4
• An overall commitment to quality and emphasis on using documented processes is also a very important
COMSATS Institute of Information Technology, Attock
20 Farhan Aadil
20
COMSATS Institute of Information Technology, Attock
21 Farhan Aadil
21
Inspections - 1
• Inspections, by all accounts, do a better job of error removal than any competing technology, and they do it at a lower cost
COMSATS Institute of Information Technology, Attock
22 Farhan Aadil
22
Inspections - 2
• Inspections are conducted by a group of people working on the project, with the objective to remove defects or errors
COMSATS Institute of Information Technology, Attock
23 Farhan Aadil
23
Inspections - 3
• Requirements errors detected during this inspections save lot of money and time as requirements errors do not flow into the design and development phases of software
COMSATS Institute of Information Technology, Attock
24 Farhan Aadil
24
Inspections - 4
• A complete description of inspections must address five dimensions:
COMSATS Institute of Information Technology, Attock
25 Farhan Aadil
25
Defect Detection Without Inspections
Requirements Design Coding Documentation Testing Maintenance
Requirements Design Coding Documentation Testing Maintenance
Defect Origins
COMSATS Institute of Information Technology, Attock
26 Farhan Aadil
26
Defect Detection With Inspections
Requirements Design Coding Documentation Testing Maintenance
Requirements Design Coding Documentation Testing Maintenance
Defect Origins
COMSATS Institute of Information Technology, Attock
27 Farhan Aadil
27
Observations
• Requirements engineers are trained to write requirements documents, but have no training on reading/reviewing
requirements documents
COMSATS Institute of Information Technology, Attock
28 Farhan Aadil
28
Techniques for Reading Requirements
Documents
• Ad hoc review • Checklist review
• Defect-based reading
COMSATS Institute of Information Technology, Attock
29 Farhan Aadil
29
Ad hoc Review
COMSATS Institute of Information Technology, Attock
30 Farhan Aadil
30
Checklist Review
COMSATS Institute of Information Technology, Attock
31 Farhan Aadil
31
Defect-based Reading
• Provides a set of systematic procedures that reviewers can follow, which are tailored to the formal software cost
COMSATS Institute of Information Technology, Attock
32 Farhan Aadil
32
Perspective-Based Reading (PBR)
• Researchers at Experimental Software Engineering Group at the University of Maryland, College Park, have created
Perspective-Based Reading (PBR) to provide a set of
COMSATS Institute of Information Technology, Attock
33 Farhan Aadil
33
Different Perspectives - 1
• PBR operates under the premise that different information in the requirements is more or less important for the different uses of the document
COMSATS Institute of Information Technology, Attock
34 Farhan Aadil
34
Different Perspectives - 2
• PBR provides a set of individual reviews, each from a
particular requirements user’s point of view, that collectively cover the document’s relevant aspects
COMSATS Institute of Information Technology, Attock
35 Farhan Aadil
35
Steps in PBR
• Selecting a set of perspectives for reviewing the requirements document
• Creating or tailoring procedures for each perspective usable for building a model of the relevant requirements information
• Augmenting each procedure with questions for finding defects while creating the model
COMSATS Institute of Information Technology, Attock
36 Farhan Aadil
36
Two Questions
• What information in these documents should they check?
COMSATS Institute of Information Technology, Attock
37 Farhan Aadil
37
Benefits of Different
Perspectives - 1
• Systematic
– Explicitly identifying the different uses for the
requirements gives reviewers a definite procedure for verifying whether those uses are achievable
• Focused
COMSATS Institute of Information Technology, Attock
38 Farhan Aadil
38
Benefits of Different
Perspectives - 2
• Goal-oriented and customizable
– Reviewers can tailor perspectives based on the current goals of the organization
• Transferable via training
– PBR works from a definite procedure, and not the
COMSATS Institute of Information Technology, Attock
39 Farhan Aadil
39
Identifying Defects
• A series of questions are used to identify different types of requirements defects
• Requirements that do not provide enough information to answer the questions usually do not provide enough
COMSATS Institute of Information Technology, Attock
40 Farhan Aadil
40
Requirements Defects that PBR Helps
Detect
• Missing information • Ambiguous information • Inconsistent information • Incorrect fact
COMSATS Institute of Information Technology, Attock
41 Farhan Aadil
41
Missing Information - 1
• Any significant requirement related to functionality,
performance, design constraints, attributes, or external interface not included
COMSATS Institute of Information Technology, Attock
42 Farhan Aadil
42
Missing Information - 2
• Sections of the requirements document
COMSATS Institute of Information Technology, Attock
43 Farhan Aadil
43
Ambiguous Information
COMSATS Institute of Information Technology, Attock
44 Farhan Aadil
44
Inconsistent Information
COMSATS Institute of Information Technology, Attock
45 Farhan Aadil
45
Incorrect Facts
COMSATS Institute of Information Technology, Attock
46 Farhan Aadil
46
Extraneous Information
COMSATS Institute of Information Technology, Attock
47 Farhan Aadil
47
Miscellaneous Defects
COMSATS Institute of Information Technology, Attock
48 Farhan Aadil
48
Benefits of PBR’s Detailed Questions
• Allow controlled improvement
– Reviewers can reword, add, or delete specific questions
• Allow training
– Reviewers can train to better understand the parts of a
COMSATS Institute of Information Technology, Attock
49 Farhan Aadil
49
PBR General Questions - 1
• Does the requirement make sense from what you know
about the application or from what is specified in the general description?
• Do you have all the information necessary to identify the inputs to the requirement? Based on the general
COMSATS Institute of Information Technology, Attock
50 Farhan Aadil
50
PBR General Questions - 2
• Have any of the necessary inputs been omitted?
• Are any inputs specified that are not needed for this requirement?
COMSATS Institute of Information Technology, Attock
51 Farhan Aadil
51
Results of PBR Experiments - 1
• PBR provides a framework that represents an improved approach for conducting requirements reviews
COMSATS Institute of Information Technology, Attock
52 Farhan Aadil
52
Results of PBR Experiments - 2
• PBR seems best suited for reviewers with a certain range of experience (not too little; not too much)
COMSATS Institute of Information Technology, Attock
53 Farhan Aadil
53
Results of PBR Experiments - 3
• Relatively novice reviewers can use PBR techniques to apply their expertise in other development tasks to defect detection
COMSATS Institute of Information Technology, Attock
54 Farhan Aadil
54
Results of PBR Experiments - 4
• It creates high-level representations of the software system, usable as a basis of work products in later stages of the
development
COMSATS Institute of Information Technology, Attock
55 Farhan Aadil
55
Results of PBR Experiments - 5
• PBR facilitates controlled improvements, providing a definite procedure, alterable according to projects metrics and
COMSATS Institute of Information Technology, Attock
56 Farhan Aadil
56
Summary
• Introduced the concept of requirements errors and types of requirements errors
• Discussed the impact of requirements errors • Discussed error prevention in requirements
COMSATS Institute of Information Technology, Attock
57 Farhan Aadil
57
References
• ‘Requirements Engineering: Processes and Techniques’ by G. Kotonya and I. Sommerville, John Wiley & Sons, 1998
• Software Quality: Analysis and Guidelines for Success by Capers Jones
• ‘How Perspective-Based Reading Can Improve
COMSATS Institute of Information Technology, Attock
58 Farhan Aadil
58
References - 2
• Software Engineering: A Practitioner’s Approach by Roger S. Pressman
• A Handbook of Software Quality Assurance edited by G. Gordon Schulmeyer and James L. McManus
• Customer-Oriented Software Quality Assurance by Frank P. Ginac