Committees responsible
Foreword
Section 1. General
0 Introduction
1 Scope
2 Definitions and mathematical notation
2.1 Definitions
2.2 Mathematical notation
Section 2. Basic considerations
3 Software and reliability
4 Collection of reliability data
5 Introduction to software reliability assessment
      methods
5.1 Sources of unreliability
5.2 Use of statistical methods
5.3 Objectives
5.4 Principal methods
6 Metrics and model criteria
6.1 Output parameters
6.2 Input parameters
6.3 Criteria for evaluating software reliability
      models
Section 3. Procedures
7 Reliability assessment by software developers
7.1 Assessment activities and uses
7.2 Baseline
7.3 Process properties
7.4 Product properties
7.5 Failure data
7.6 Fault data
7.7 Processing and storing failure reports
7.8 Uses of reliability assessment
7.9 Preliminary analysis of data
8 Data collection forms
8.1 General guidance
8.2 Form 1: incident report
8.3 Form 2: fault report
8.4 Form 3: change report
8.5 Form 4: software part report
8.6 Form 5: process definition report
8.7 Form 6: software use log (calendar time)
8.8 Form 7: software use log (usage time)
9 Role of a central database
9.1 Concept
9.2 Communication with the central database
9.3 Functions and facilities
Section 4. Software reliability assessment modelling
10 Classification and description of models
10.1 Criteria for classification
10.2 Failure data models
10.3 Properties models
10.4 Summary of models
11 Introduction to failure data models
11.1 The fundamental problem
11.2 Prediction method
11.3 Classification of failure data models
12 Exponential order statistic (EOS) models
12.1 EOS model family
12.2 Assumptions
12.3 Definitions
12.4 Deterministic EOS models
12.5 Doubly-stochastic EOS (DS/ESO) models
12.6 Doubly-stochastic EOS IIDOS models
12.7 Doubly-stochastic EOS NHPP models
12.8 Doubly-stochastic EOS distribution-free models
13 Inter-failure time models
13.1 Family characteristics
13.2 Littlewood-Verrall (L-V)
13.3 Inter-failure time models other than Littlewood-
      Verrall
14 Proportional hazard model
15 Non-parametric analysis
15.1 Basis
15.2 Miller and Sofer
15.3 Isotonic regression variant
15.4 Limitations
15.5 Example
16 Structural models
16.1 Littlewood structural model
16.2 Shooman structural model
17 Miscellaneous failure data models
17.1 Overview
17.2 Input domain based models
17.3 Seeding and tagging
17.4 Time series analysis
17.5 Availability models
17.6 High reliability assessment
17.7 General reliability growth model applications
18 Inference procedures for failure data models
18.1 Initial considerations
18.2 Search procedure
18.3 Objective function
18.4 Bayesian estimation
18.5 Other methods
19 Procedures for assessment and manipulation of
      failure data models
19.1 Criteria for comparison
19.2 u-plots
19.3 y-plots
19.4 Prequential likelihood
19.5 Adaptive modelling
19.6 Combination of model predictions
20 Software properties and development process
      properties models
20.1 General considerations
20.2 Software properties models
20.3 Software development process models
20.4 Combined approaches
21 Recommendations on the application of reliability
      assessment methods
21.1 Basic criteria
21.2 Selection of failure data models
21.3 Software and development process properties models
Appendices
A Bibliography
B Software Data Library
Table
1 Summary of some reliability models
Figures
1 Error, fault, failure relationship
2 Software data feedback
3 Flow of failure data information
4 Example of failure history graphs and use of LCM
5 Graph of failure rate against calendar time
6 Graph of failure rate against execution time
7 Form 1: incident report
8 Form 2: fault report
9 Form 3: change report
10 Form 4: software part report
11 Form 5: process definition report
12 Form 6: software use log (calendar time)
13 Form 7: software use log (usage time)
14 Fundamental reliability assessment problem: time
      to failure data
15 Fundamental reliability assessment problem:
      failure count data
16 Family tree of failure data models
17 Illustration of why the assumption of uniform
      fault size leads to optimistic estimates
18 Observed number of failures as a function of time:
      n(t) = number of failures in [0,t]
19 Piecewise-linear version of counting function from
      figure 18. Simulated process is NHPP:
      r(t) = 0.2228 [(t/100) to the power -0.443]
20 Piecewises-constant empirical rate functions with
      constant interval width for data from figure 18.
      True failure rate function is superimposed:
      r(t) = 0.2228 [(t/100) to the power -0.443]
21 Estimate of the failure rate function for data in
      figure 18 using d = 1. True failure rate function
      is superimposed
22 Estimate of failure rate function for data in
      figure 18 using d = 2. True failure rate function
      is superimposed
23 Estimate of failure rate function for data in
      figure 18 using d = 3. True failure rate function
      is superimposed
24 Example of a u-plot: assessment of bias in
      predictions