10. PLANEACIÓN
11.1 INFORMACIÓN DOCUMENTADA DEL SGSI TIC
Quality control includes many general tools and techniques. We will discuss some of the important ones, i.e., Pareto analysis, statistical sampling, Six Sigma, and quality control charts, and how these techniques can be applied to software projects.
9.5.1
Pareto analysis
Pareto analysisis used to identify the vital few contributors that account for most quality problems in a system. It is also referred to as the 80-20 rule, meaning that 80% of problems are often due to 20% of the causes.
Pareto diagrams are histograms, or column charts representing a frequency distribu- tion that help identify and prioritize problem areas. The variables described by the histogram are ordered by frequency of occurrence. Therefore, Pareto diagrams help visualise which factors contribute most to quality problems.
9.5.2
Statistical sampling
Statistics play an important role in project quality management. Important concepts in statistics that are useful in quality management include statistical sampling, certainty factor, standard deviation, and variability.
Statistical sampling involves choosing part of a population of interest for inspection. The sample sizeis an important factor in determining how representative the sample is. A simple formula for determining sample size is:
S = 0.25× F E 2 , (9.1)
whereS is the sample size,F is thecertainty factor, andE is theacceptable error. Thecertainty factor denotes how certain one wants to be that the data sampled will not include variations that do not naturally exist in the population. There is a one-to-one mapping between the certainty factor and desired certainty,e.g., 95% desired certainty corresponds to the certainty factor of 1.96. The students should refer to Table 8-2 in the textbook for some commonly used certainty factors. Note that the acceptable error equals 100% minus the desired certainty.
9.5.3
Six sigma
Six Sigmais defined as ‘a comprehensive and flexible system for achieving, sustaining and maximizing business success’ by Pande et al. in their bookThe six sigma way. Software projects that use Six Sigma principles for quality control normally follow a five-phase improvement process called DMAIC, which stands for Define, Measure, Analysse, Improve, and Control:
Ê
Define: Define the problem/opportunity, process, and customer requirements.Ë
Measure: Define measures, collect, compile, and display data.Ì
Analyse: Scrutinize process details to find improvement opportunities.Í
Improve: Generate solutions and ideas for improving the problem.Î
Control: Track and verify the stability of the improvements and the predictability of the solution.Six Sigma principles differ from previous quality control initiatives such astotal quality management (TQM) andbusiness process reengineering(BPR) in having several new ideas:
Ê
using Six Sigma principles is an organisation-wide commitmentË
organisations that successfully implement Six Sigma principles have the ability and willingness to adopt two seemingly contrary objective at the same timeÌ
Six Sigma is not just a program or a discipline to organisations, but anoperating philosophy that is customer-focused and strives to drive out waste, raise levels of quality, and improve financial performance atbreakthroughlevels.Organisations that adopt Six Sigma principles set high goals and use the DMAIC im- provement process to achieve extraordinary quality improvements.
Project selection is especially important for Six Sigma projects. Not all organisations can benefit from Six Sigma because minimising is not meaningful if an organisation is making a product that has no market. Three criteria exist for potential Six Sigma projects:
Ê
There must be a quality problem or gap between the current and desired perfor- mance.Ë
Six Sigma projects are also best suited for projects that do not have a clearly understood problem.Six Sigma projects can be understood as simple types of projects that focus on sup- porting the Six Sigma philosophy by being customer-focused and striving to drive out waste, raise levels of quality, and improve financial performance at breakthrough levels. The termSigmameans standard deviation, which measures how much variation exists in a distribution of data, and is represented by the Greek symbol σ. Standard devi- ation is a key factor in determining the acceptable number of defective units found in a population. In pure statistical terms, being±6σ means 99.9999998% of population within range, or onlytwo defective units per billion. However, there is a1.5σ shift due to disagreement among statistical experts regarding how Six Sigma measures are de- fined. The result is that the target for Six Sigma programs is 3.4 defects per million
opportunities, or 99.99966% of population within range.
Six 9s of quality is another term used in the telecommunications industry. It is a mea- sure of quality control equal to 1 fault in 1 million opportunities. It means 99.9999% service availability or 30 seconds of downtime a year.
9.5.4
Quality control charts and the seven run rule
Quality control charts are graphical display of data that illustrates the results of a pro- cess over time. Such a chart allows project managers to determine whether a process is in control or out of control. For a process in control, any variations in the results of the process are caused by random events. On the other hand, for a process out of control, variations are caused bynonrandomevents. The causes resulting nonrandom events need to be identified and eliminated through adjusting the process.
The seven run rule states that if seven data points in a row are all below the mean, above the mean, or are all increasing or decreasing, then the process needs to be examined for nonrandom events. The rule can be utilised to identify abnormal data patterns in quality control charts.
9.5.5
Testing
Software projects use testing extensively to ensure quality. In a software system devel- opment life cycle, testing needs to be done during almost every phase of the cycle.
Unit testing is to test each individual component (often a program) to ensure it is as defect-free as possible. Integration testingis to test functionally grouped components. It ensures a subset(s) of the entire system works together. System testing tests the entire system as one entity.User acceptance testing is an independent test performed by end-users prior to accepting the delivered system.
Figure 9.2: Testing Tasks in the Software Systems Development Life Cycle (source: Schwalbe 2006, Figure 8-4)