Copyright © 1994 Carnegie Mellon University1Disciplined SoftwareEngineeringLecture #6 Software Engineering Institute Carnegie Mellon University Pittsburgh, PA 15213 Sponsored by the U.S. Department of DefenseCopyright © 1994 Carnegie Mellon University2Process Measurement Principles of process measurement The goal-question-metric (GQM) paradigm Basic process measures Gathering and using data The process baselineCopyright © 1994 Carnegie Mellon University3Process Measurement Principles To be useful, measurements should be•gathered for a specific purpose•explicitly defined•properly managed•properly used Measuring your process will not improve it.You must make process changes to achievelasting improvement.Copyright © 1994 Carnegie Mellon University4Process Measurement Purposes We measure for the following reasons•to understand and manage change•to predict or plan for the future•to compare one product, process, ororganization with another•to determine adherence to standards•to provide a basis for controlCopyright © 1994 Carnegie Mellon University5Measurement Objectives Measurements only produce numbers. To be useful, they must•relate to business objectives•be properly interpreted•lead to appropriate action If the business purposes for the measurementsare not understood•the wrong data may be gathered•the data may not be properly usedCopyright © 1994 Carnegie Mellon University6Types of Measurements - 1 Measurements may be•objective/subjective•absolute/relative•explicit/derived•dynamic/static•predictive/explanatory We generally seek objective, absolute, andexplicit measures They may be dynamic or static, predictive orexplanatoryCopyright © 1994 Carnegie Mellon University7Types of Measurements - 2 We seek useful relationships that correlate, forexample•program size versus development hours•cost distributions•defect densities We also seek controlling or predictiveparameters, for example•actions to reduce test defects•steps to improve review quality•means to improve productivityCopyright © 1994 Carnegie Mellon University8The Goal-Question-Metric (GQM)Paradigm The GQM paradigm establishes a framework forgathering and using data. It starts with an explicit statement of datagathering goals Next it defines the questions the data are toanswer And then defines the data metricsCopyright © 1994 Carnegie Mellon University9Measurement Goals What are the goals for which data are required?•these may be personal, project, or business•they should be explicit Be clear on whose goals these are•yours, your project’s, your management’s•try to understand the goals at all these levels Relating your goals to other’s goals provides•a clearer context for the measurements•more likely support for your workCopyright © 1994 Carnegie Mellon University10Measurement Questions What will it take to meet these goals?•plan the actions required•who must take these actions? Do people need to be convinced to act?•are data needed to convince them?•how will these data be used? What is needed to implement the actions?•what data are required?•how will these data be used?Copyright © 1994 Carnegie Mellon University11Measurement Metrics Precisely what data are needed?•define the data so others could gather them•establish specific fields and formats How will these data be gathered?•provide data gathering forms•define the data gathering process How will these data be retained and used?•define the specific calculations/analyses•work through some test casesCopyright © 1994 Carnegie Mellon University12Making the Measurements You are now ready to start the data gatheringprocess First•communicate the entire GQM framework•ensure that the process is understood•conduct a test if necessary Start data gathering•monitor data quality•provide feedbackCopyright © 1994 Carnegie Mellon University13GQM for the PSP The PSP has the basic goal of assisting you inimproving your personal performance This suggests some likely personal goals to•understand your personal process•determine steps to improve product quality•understand your personal productivity•establish benchmarks for measuringimprovement•make better plansCopyright © 1994 Carnegie Mellon University14Some General PSP Questions How good is my process? Where can it be improved? What is the most important improvement Ishould make now? What are others doing that works better? How can I learn from them?Copyright © 1994 Carnegie Mellon University15How Good is my Process? The principal dimensions of process quality are•product quality•predictability•productivity•cycle time Since improvement must start from the currentprocess, the first step is to establish measuresof current process quality. You can then ask more informed questions.Copyright © 1994 Carnegie Mellon University16Measuring Process Quality Start with building a basic understanding ofyour process•what do you do?•how much time do you spend?•how is this time distributed?•how predictable is your work?•how stable is your work?•what is the quality of the products youproduce? The PSP addresses these issues by gatheringbasic process dataCopyright © 1994 Carnegie Mellon University17The PSP Measurements The basic PSP data are•program size•time spent by phase•defects found and injected by phase Both actual and estimated data are gathered onevery item The derived measures then establish•data to support planning•measures of process qualityCopyright © 1994 Carnegie Mellon University18PSP Size Measures The goals of the PSP size measures are to•define a consistent size measure•establish a basis for normalizing time anddefect data•help make better size estimates The questions asked are•what size program did I plan to develop?•how good was my size estimate?•what was the complete size description of thefinished program?Copyright © 1994 Carnegie Mellon University19PSP Time Measures The goals of the PSP time measures are to•determine how much time you spend in eachPSP phase•help you to make better time estimates The questions asked are•how much time did you spend by PSP phase?•how much time did you plan to spend by PSPphase?Copyright © 1994 Carnegie Mellon University20PSP Defect Measures The goals of the PSP defect measures are to•provide an historical baseline of defect data•understand the numbers and types of defectsinjected•understand the relative
View Full Document