๐ Lecture 27 โ Software Process Improvement Using PDCA
๐ Overview: This lecture introduces the PDCA (Plan-Do-Check-Act) cycle, also known as the Shewart cycle, as a structured and cyclic model for implementing software process improvement (SPI) programs in organizations. It explains the four phases in detail and presents a spiral model that combines PDCA with the classic spiral process model, illustrating a staged progression from initial project-level adoption to widespread organizational use. The lecture emphasizes critical aspects such as planning for risks, structuring for success by identifying key people, measuring success, and leveraging achievements to ensure lasting improvement.
๐๏ธ Topics Covered
The lecture covers the Shewart/PDCA cycle and its combination with the spiral model for process improvement adoption across five cycles (from first-time practice usage to optimization). It then details the Plan quadrant, including identifying risks, business planning, core-competence planning, and management commitment influencers. Next, it examines the Do quadrant, focusing on structuring for success and the roles of key people like project managers, champions, sponsors, opinion leaders, and infrastructure members. The Check quadrant discusses measuring success using primitive metrics, and the Act quadrant covers leveraging success through revision, spreading improved processes, and summarizing the best way to extend adoption.
๐ Lecture Summary
Software Process Improvement Using PDCA
There are a number of models for instituting software process improvement (SPI) programs in organizations. All these models have to be incorporated in an orderly and cyclic manner, and these introductions cannot and must not be made abruptly.
Shewart cycle
The Shewart cycle is one such way of introducing a software process improvement program in an organization. It was popularized by W. Edwards Deming. The Shewart cycle has four steps: Plan, Do, Check, Act (PDCA). The PDCA model can be combined with the famous spiral process model, specially adapted for process improvements, resulting in a new spiral process improvement model for software.
Cycle I: SPI Spiral Model
The project plans a new practice (Plan), tries the new practice (Do), achieves success (Check), and then describes the practice as part of that success (Act).
Cycle II: SPI Spiral Model
Other projects plan to use the practice (Plan), with minimal new documentation added (Do). Other successes are linked to the practice (Check), and a decision for "organization-wide" use is made, with successes told widely (Act).
Cycle III: SPI Spiral Model
People are assigned to document, train, and support (Plan). A training manual, process support by staff, and initial process metrics are created (Do). More projects use the practice with few hard problems, and initial measurable results emerge (Check). Procedures are revised, expected results are quantified, and management refers to the practice as a "standard" (Act).
Cycle IV: SPI Spiral Model
All projects plan to use the practice, with urgent requests for training and improvements planned (Plan). Support processes are standardized, initial automation is implemented, and metrics are used to monitor effectiveness (Do). Most projects use the practice, which is evaluated against measurable goals, and most people are convinced to value it (Check). Procedures are revised, ongoing responsibility for effective use is assigned, and management reviews effectiveness (Act).
Cycle V: SPI Spiral Model
Plans are made to optimize the infrastructure (Plan). The number of infrastructure personnel is reduced to optimal ongoing levels (Do). Metrics are monitored to ensure no loss of benefits (Check). Effectiveness is reviewed, and the process is tuned (Act).
The figure shows a progression from first-time practice usage (in the center) to widespread organizational adoption. The model includes several patterns across the plan/do/check/act quadrants, reflecting increasing people and resource investments, greater management understanding, attention, and commitment, and better control with process metrics.
Plan
Most new improvements are first tried by one project team. If successful, these improvements are spread to other projects. As usage spreads, planning aspects include ways to provide better documentation, training, and support. Eventually, demand outpaces the ability of the people helping new adopters, and plans for an ongoing infrastructure are developed and tuned.
Plan: Identifying Risks and Deciding What to Do
While PDCA always starts with some form of commitment, the nature and strength of that commitment is constantly challenged as plans solidify. Business planning includes changes that make good business sense and link to your goals. You need to plan the scope, timing, cost, and payback of each change. Without such scoping, you have no basis for comparing proposed process-improvement projects to any other business proposals. You must understand your organization's readiness and enthusiasm for each proposed change. Your culture, your existing processes, and your people's underlying beliefs all affect changes. Focus on developing core-competence is extremely important for organizations in the software business. Core-competence planning is just as important as business planning in setting an organization's long-term strategy. The key results of core-competence planning are a set of clearly stated goals and action plans with clear links to business needs. An investment model of your software processes will help you to evaluate trade-offs. No matter how well you analyze the business urgency of a particular process improvement and its investment value, your organization might just not be ready for it. A good subjective assessment should give you some sense of how widespread feelings are about key issue areas and how motivated people are to change.
A table shows Management Commitment Influencers which include Business (Strategic Vision, Business Approach, Core Competence) and Tactical (Customer Perceptions, Market Share, Product Cycle Time, Profitability) aspects, and Organizational (Organizational Maturity, Process Improvement Infrastructure) and Tactical (Organizational Inertia, Stability, Cost/Time Alignment) aspects. When these are not positive, they can distract and undermine your manager's best intentions. Strategic business aspects most reflect the leadership of your organization's management. Tactical business aspects are a constant source of tension, tied to commonly accepted evaluations of business health. Strategic organizational aspects can be the strongest negative influencers. Tactical organizational aspects are the most controllable from lower in the organization. A strong selection process with a well-thought-out balance of tactical and strategic projects with clearly defined goals shows people your rationale and helps to motivate them toward successful change.
Do
Progress in this quadrant parallels the stages in the Plan quadrant. Adoption by second-ring projects starts a trend to improve and adapt documentation and training. By the third ring, users have formed realistic expectations for the new process and use metrics. In the outermost two rings, good support is common, automation improves process efficiency and quality, and the number of infrastructure personnel is finally reduced to optimal ongoing levels.
Do: Structuring for Success
Software managers often respond to proposed process-improvement projects with skepticism. One of the most powerful approaches is to work with people to mentally shift to a future desired state. The spiral model shows that improvements progress by stages. One aspect is the key people in an organization that require special attention.
Project Managers play a key role by ensuring that all involved people work together productively. A project manager is generally also the project's champion. A sponsor is a person in a higher organizational position who ensures funding and high-level public enthusiasm. Sometimes there are less obvious technical leaders, called opinion leaders. Before other people will consider a change, they look to whether opinion leaders accept it. When people first work with a new process, ready access to infrastructure members to answer questions and remove roadblocks is invaluable. The single most important step you can take is to create and manage from a good process-improvement project plan. A set of storyboard pictures or maps of where you plan to go is also a very useful way to start a project plan. A typical storyboard outline includes sections for Situation, Challenges, Plan (Project Selection Criteria, Planned Changes, Objectives, Cost/Benefit Analysis, Project Milestones), Do (Methods for Measuring and Validating), Check (Approach for Standardizing Process Changes), and Act.
๐ Storyboard: A visual outline or map used to organize thoughts for a process-improvement project plan, often used to create a slideset that evolves over the adoption of the process improvement.
Check
After the first ring project team uses a new practice, they describe their "success" in terms matching whatever measurements they took. As other projects adopt the new process in later spiral rings, improvement metrics become more precise until process performance becomes well understood and is measured regularly to ensure optimum performance is maintained.
Check: Measuring Success
If you started with a storyboard that included the Check part, you are probably in good shape. The amount of effort needed to set up measures depends on what measures already exist. If you do not measure these four primitive metrics, it will be difficult to frame expectations and even more difficult to describe success.
- Non-Comment Source Statements (NCSS)
- Engineering Months (EM)
- Defects
- Calendar Months (CM)
๐ Primitive Metrics (NCSS, EM, Defects, CM): Four fundamental measurements (Non-Comment Source Statements, Engineering Months, Defects, Calendar Months) that are essential for framing expectations and describing success for a software process improvement project.
Act
The speed at which adoptions of a new practice occur will heavily depend on how credible early adopters are, how good a job they did of measuring their processes, how well they can communicate their success, and how well their improvement matches the organization's readiness. Gaining ever-increasing management commitment as you move outward on these segments suggests the critical elements of follow-through and "sales" that are necessary for process improvements.
Act: Leveraging Success
The essence of the "Act" step is to revise and spread improved processes. This is done by describing successes so different audiences understand and get excited. No process should ever be "final". This means we must model improvements on previous successes, create similar situations and environments, and reinforce behaviors that move people as quickly as possible to improved conditions. The best way to summarize what is needed to extend best-practice adoption and effective usage outward is:
- Proactively identify and seek support of process-improvement champions and sponsors.
- Reinforce management awareness and commitment with a strong business case.
- Build an infrastructure strong enough to achieve and hold software core competence.
- Measure the extent of adoption of each desired process improvement until it is operating effectively, efficiently, and across all appropriate parts of the organization.
Conclusions
Use the PDCA model to help you make sure you have clear goals and plans for what you want at the end of each of multiple spirals, understand how much you can expect to successfully do in stages, measure progress and results that emphasize your incremental successes, and evaluate management and organizational commitment influencers and regularly reenlist and reinforce support on your continued exciting journey.
โญ Key Takeaways
The PDCA (or Shewart) cycle is a structured, four-phase approach to introducing software process improvements: Plan (identify risks and decide), Do (train, adapt, and structure for success), Check (evaluate results and measure success using primitive metrics like NCSS, EM, Defects, and CM), and Act (revise, spread, and leverage success). The spiral model combines PDCA with a progressive, cyclic framework, moving from a single project trying a new practice in the center to widespread organizational adoption and optimization at the outer rings, reflecting increasing investment, management commitment, and process control. Success in the "plan" phase requires careful risk identification, business and core-competence planning, and evaluation of management commitment influencers, while the "do" phase hinges on identifying and leveraging key people (project managers, champions, sponsors, opinion leaders, infrastructure members) and using storyboards for effective project planning. The end goal is to shift thinking from solely tactical process-improvement projects to strategically oriented organizational capabilities that evolve key abilities from individual knowledge to robust organizational infrastructure.
๐ง Quick Revision Questions
- What are the four steps of the Shewart (PDCA) cycle, and what is the primary action that must be taken in each step according to the lecture?
- Describe the progression of adoption from Cycle I to Cycle V in the spiral model for process-improvement adoption.
- List the four primitive metrics (the minimum measurements) required to frame expectations and describe success for a process improvement project.
- What are the four types of key people identified in the "Do: Structuring for Success" section, and what is the primary role of a "sponsor" compared to a "champion"?
- Name the twelve management commitment influencers listed in the table, and categorize them into the four main groups (Business/Strategic, Business/Tactical, Organizational/Strategic, Organizational/Tactical).
๐ Lecture 28 โ Software Process Improvement Using SEIโs IDEAL Model
๐ Overview: This lecture introduces the IDEAL model, a structured organizational improvement framework developed by the SEI. It serves as a roadmap for initiating, planning, and implementing improvement actions in software process improvement. The model provides a disciplined engineering approach for continuous improvement, focusing on managing the improvement program and establishing a long-term improvement strategy.
๐๏ธ Topics Covered
The lecture covers the IDEAL model's five phases: Initiating, Diagnosing, Establishing, Acting, and Learning. For each phase, detailed activities are explained. The initiating phase covers stimulus for change, setting context, building sponsorship, and chartering infrastructure. The diagnosing phase covers characterizing current and desired states and developing recommendations. The establishing phase covers setting priorities, developing approach, and planning actions. The acting phase covers creating solution, piloting/testing, refining, and implementing solution. The learning phase covers analyzing/validating and proposing future actions. Useful tips on SPI are also provided.
๐ Lecture Summary
Introduction
The IDEAL model is an organizational improvement model that serves as a roadmap for initiating, planning, and implementing improvement actions. It is named for the five phases it describes: initiating, diagnosing, establishing, acting, and learning. As originally conceived, it was a life-cycle model for software process improvement based upon the Capability Maturity Modelยฎ (CMMยฎ) for Software, but the SEI has revised it for broader application. The model provides a usable, understandable approach to continuous improvement by outlining the steps necessary to establish a successful improvement program.
๐ก Why this matters: Following the phases, activities, and principles of the IDEAL model has proven beneficial in many improvement efforts, providing a disciplined engineering approach that focuses on managing the improvement program and establishes the foundation for a long-term improvement strategy.
๐ Definition โ IDEAL Model: An organizational improvement model that serves as a roadmap for initiating, planning, and implementing improvement actions, named for its five phases: Initiating, Diagnosing, Establishing, Acting, and Learning.
The Initiating Phase
Critical groundwork is completed during the initiating phase. The business reasons for undertaking the effort are clearly articulated. The effort's contributions to business goals and objectives are identified, as are its relationships with the organization's other work. The support of critical managers is secured, and resources are allocated on an order-of-magnitude basis. Finally, an infrastructure for managing implementation details is put in place.
๐ก Why this matters: If the initiating phase activities are done completely and well, subsequent activities can proceed with minimal disruption. If done poorly, incompletely, or haphazardly, then time, effort, and resources will be wasted in subsequent phases.
Stimulus for Change: It is important to recognize the business reasons for changing an organization's practices. The stimulus for change could be unanticipated events or circumstances, an edict from someone higher up in the organization, or the information gained from benchmarking activities as part of a continuous improvement approach. Whatever the stimulus, it can have far-reaching influence on the effort's visibility, conduct, and ultimate success. Change for the sake of change rarely results in significant improvement. When the business reasons for change are more evident, there is greater buyin throughout the organization and greater chances for success.
Set Context: Once the reasons for initiating change have been clearly identified, the organization's management can set the context for the work. "Setting context" means being very clear about where this effort fits within the organization's business strategy. Questions to answer include: What specific business goals and objectives will be realized or supported by this change? How will it affect other initiatives and ongoing work? What benefits (such as return on investment or improved capabilities and morale) will result? Context and implications often become more evident as the effort proceeds, but it is important to be as clear as possible regarding these issues early in the effort.
Build Sponsorship: Effective sponsorship is one of the most important factors for improvement efforts. It is necessary to maintain sponsorship levels throughout an improvement effort, but because of the uncertainty and chaos facing the organization in the beginning, it is especially important to build critical management support early in the process. The commitment of essential resources is an important element of sponsorship, but effective sponsors often do much more than this. Sponsors can be most effective if they give personal attention to the effort and stick with it through difficult times.
Charter Infrastructure: Once the reason for the change and the context are understood and key sponsors are committed, the organization must set up a mechanism for managing the implementation details. The infrastructure may be temporary or permanent, and its size and complexity may vary substantially depending on the nature of the improvement. For a small effort, the infrastructure may be a single part-time employee; for a large and complex effort, such as software process improvement, it may involve 2-3% of the organization's people across a number of groups. Chartering the infrastructure involves developing explicit written agreements that document and clarify expectations and describe responsibilities.
The Diagnosing Phase
The diagnosing phase builds upon the initiating phase to develop a more complete understanding of the improvement work. During this phase, two characterizations of the organization are developed: the current state of the organization and the desired future state. These organizational states are used to develop an approach for improving business practice.
Characterize Current and Desired States: Characterizing the current and desired states is similar to identifying the origin and destination of a journey. This can be done more easily using a reference standard such as the CMM for Software. Where such a standard is not available, a good starting point is the factors identified as part of the "stimulus for change" activity. This activity should focus on elements critical to the changes being introduced, rather than every aspect of an organization's work.
Develop Recommendations: The recommendations developed as a part of this activity suggest a way of proceeding in subsequent activities. The diagnosing phase activities are most often performed by a team with experience and expertise relevant to the task at hand. Their recommendations often weigh heavily in the decisions made by key managers and sponsors.
The Establishing Phase
The purpose of the establishing phase is to develop a detailed work plan. Priorities are set that reflect the recommendations made during the diagnosing phase as well as the organization's broader operations and the constraints of its operating environment. An approach is then developed that honors and factors in the priorities. Finally, specific actions, milestones, deliverables, and responsibilities are incorporated into an action plan.
Set Priorities: The first activity of this phase is to set priorities for the change effort. These priorities must take many factors into account: resources are limited, dependencies exist between recommended activities, external factors may intervene, and the organization's more global priorities must be honored.
Develop Approach: Combining increased understanding of the scope of work (gained in the diagnosing phase) with a set of priorities leads to the development of a strategy for accomplishing the work and identifying resource availability. Technical factors might include the specifics of installing the new technology and new skills and knowledge required for using a technology. Non-technical factors, including the organization's culture, likely sources of resistance, sponsorship levels, and market forces, also must be considered.
Plan Actions: With the approach defined, a detailed implementation plan can be developed. This plan includes schedule, tasks, milestones, decision points, resources, responsibilities, measurement, tracking mechanisms, risks and mitigation strategies, and any other elements required by the organization.
The Acting Phase
The activities of the acting phase help an organization implement the work that has been conceptualized and planned in the previous three phases. These activities will typically consume more calendar time and more resources than all of the other phases combined.
Create Solution: The acting phase begins with bringing all available key elements together to create a "best guess" solution to address the previously identified organizational needs. These key elements might include existing tools, processes, knowledge, and skills, as well as new knowledge, information, and outside help. The solution, which may be quite complex and multi-faceted, is often created by a technical working group.
Pilot/Test Solution: Once a solution has been created, it must be tested, as best guess solutions rarely work exactly as planned. This is often accomplished through a pilot test, but other means may be used.
Refine Solution: Once the paper solution has been tested, it should be modified to reflect the knowledge, experience, and lessons gained from the test. Several iterations of the test-refine process may be necessary to reach a satisfactory solution. A solution should be workable before it is implemented, but waiting for a "perfect" solution may unnecessarily delay the implementation.
Implement Solution: Once the solution is workable, it can be implemented throughout the organization. Various roll-out approaches may be used for implementation, including top-down (starting at the highest level and working down) and just-in-time (implementing project-by-project at an appropriate time in its life cycle). No one roll-out approach is universally better than another; the approach should be chosen based on the nature of the improvement and organizational circumstances. For a major change, implementation may require substantial time and resources.
The Learning Phase
The learning phase completes the improvement cycle. One of the goals of the IDEAL Model is to continuously improve the ability to implement change. In the learning phase, the entire IDEAL experience is reviewed to determine what was accomplished, whether the effort accomplished the intended goals, and how the organization can implement change more effectively and/or efficiently in the future. Records must be kept throughout the IDEAL cycle with this phase in mind.
Analyze and Validate: This activity answers several questions: In what ways did the effort accomplish its intended purpose? What worked well? What could be done more effectively or efficiently? Lessons are collected, analyzed, summarized, and documented. The business needs identified during the initiating phase are reexamined to see if they have been met.
Propose Future Actions: During this activity, recommendations based on analysis and validation are developed and documented. Proposals for improving future change implementations are provided to appropriate levels of management for consideration.
Useful Tips on SPI
- Proactively identify and seek support of process-improvement champions and sponsor
- Reinforce management awareness and commitment with a strong business case for each desired process improvement
- Build an infrastructure strong enough to achieve and hold software core competence
- Measure the extent of adoption of each desired process improvement until it is effectively, efficiently, and across all appropriate parts of the organization
Conclusions
The IDEAL model provides an effective approach to adopting improved software engineering processes, methods, and tools.
โญ Key Takeaways
The IDEAL model is a five-phase organizational improvement roadmap (Initiating, Diagnosing, Establishing, Acting, Learning) that provides a disciplined engineering approach to software process improvement. The initiating phase is critical as it lays the groundwork and builds sponsorship; if done poorly, resources will be wasted in subsequent phases. Effective sponsorship is one of the most important factors for improvement success, and sponsors must give personal attention and stick with the effort through difficult times. The acting phase consumes the most calendar time and resources, and solutions should be tested and refined before full implementation, with roll-out approaches chosen based on organizational circumstances. The learning phase completes the cycle by analyzing what was accomplished and proposing future actions, ensuring continuous improvement in the ability to implement change.
๐ง Quick Revision Questions
- What are the five phases of the IDEAL model, and what is the primary purpose of each phase?
- Why is the initiating phase considered so critical to the success of an improvement effort?
- What are the two characterizations developed during the diagnosing phase, and what is their purpose?
- What factors must be considered when setting priorities during the establishing phase?
- What three activities are involved in the acting phase before a solution is fully implemented, and why is each important?
๐ Lecture 29 โ Software Process Assessments
๐ Overview: This lecture introduces software process assessments, their purpose, and various methodologies used worldwide. It explains how assessments serve as diagnostic tools to identify strengths and weaknesses in software development practices, and outlines key process areas and personnel-related topics examined during assessments. The lecture also covers major assessment methodologies including SEI and SPR approaches, and presents a six-stage improvement program for organizations.
๐๏ธ Topics Covered
The lecture covers the definition and purpose of software process assessments, key process areas examined during assessments, personnel-related topics, introduction to various assessment methods including SEI and SPR approaches, and a comprehensive six-stage software improvement program covering management technologies, software processes, tools, infrastructure, reusability, and industry leadership, along with the cost, timing, and value of process improvements.
๐ Lecture Summary
Software Process Assessments
Since software applications are built by human beings, the methods, tools, and practices used for software have become subject to study and analysis. Software process assessments are normally carried out on-site by independent assessors, though in-house assessments are also common. Formal software assessments originated within the IBM corporation in the early 1970s. An assessment finds all the strengths and weaknesses associated with software.
The primary goal of software process assessments is to gather qualitative information about the practices, methods, tool suites, and organizational structures used for software. Some organizations gather additional information on topics such as office space and ergonomic factors. As software has become a worldwide business phenomenon, the need for software process assessments has become increasingly important, especially since software failures outnumber successes by a considerable margin for large systems.
Assessment data is normally gathered by means of structured interviews with managers and technical personnel. To ensure consistency among enterprises, standard questionnaires are utilized by most assessment methods, although the questions vary from method to method. Scripts and questionnaires for data collection are now utilized throughout the world.
Software Process Assessment Methodologies include:
- The SPR assessment methodology
- The SEI assessment methodology (CMMI)
- The ISO assessment methodology
- The SPICE assessment methodology
- The TickIT assessment methodology
- The Trillium assessment methodology
- The Howard Rubin Associates assessment methodology
- The Gartner Group assessment methodology
A software process assessment is somewhat equivalent to undergoing an annual medical examination. The annual physical examination is not intended to cure any specific disease โ the main purpose is to find out about the health of the patient. If the physical examination finds any serious problems, the physician can prescribe a therapy program. However, the examination itself is performed for diagnostic purposes, not for therapeutic purposes.
๐ก Why this matters: Companies often mistakenly regard the assessment as an end in itself, rather than a means to an end. Assessments by themselves will not improve software quality or productivity, but without the assessment, many problems would remain invisible. Assessments should be viewed as annual checkups to ensure progress is being made on curing problems found during the last assessment.
Key Process Areas
Out of a total of 250 factors that can affect the outcome of a software development project, an individual project would be affected by 15 to 20 factors. These factors are collected from projects in the domains of system software, information systems, military software, commercial software, outsourced projects, and end user applications.
The major topics covered by software process assessment can be summarized by ten key process areas common among many assessment methods. These areas are examined because they have a very significant impact on software project schedules, productivity, and quality levels.
Ten Key Process Areas:
- Project management methods such as estimation
- Requirements-gathering and analysis methods
- Design and specification methods
- Coding methods
- Reusability methods
- Change control methods
- User documentation methods
- Pretest defect removal methods such as inspections
- Testing methods and tools
- Maintenance methods and tools
Software process assessments attempt to identify practices that can lead to successful outcomes and seek to identify harmful practices that might lead to delays or unacceptable outcomes.
Key Personnel-Related Topics
As software projects are very labor intensive and require exceptional skills, software assessments may also examine factors concerned with hiring and supporting software personnel.
Ten Key Personnel-Related Topics:
- Staff hiring practices
- Staff training and education
- Management training and education
- Specialists and occupational groups
- Compensation levels
- Office ergonomics
- Organizational structures
- Morale surveys and results
- Work patterns and overtime
- Staff turnover rates
Introduction to Some Assessment Methods
Because software development is still a largely manual activity performed by skilled craftsmen, success on software projects requires very capable personnel led by capable managers. Good teams must utilize methods and practices known to produce excellent results and avoid practices that have led to failures.
SEI Assessment Approach
The SEI assessment data is collected by means of on-site interviews using standard questions, observations, and information data. Once collected, the assessment data are analyzed and used to place a software organization on one of the five maturity levels or capability levels of the SEI CMMI. Although the assessment and maturity-level concepts can be considered separately, most people regard the two as being a linked set.
Formal SEI assessments are performed by the SEI or by licensed consulting companies. Informal "SEI-style" assessments are performed by companies on their own or by consultants who use assessment approaches derived from published SEI materials. The SEI assessment approach originally dealt primarily with software processes and methodologies used by very large companies producing very large systems, but has now broadened the sets of factors included under the assessment.
๐ก Why this matters: Software process improvements cannot be successful if they overlook key topics or emphasize one topic and ignore others. The broad range of topics included in SEI's programs emphasizes that success in software requires being good in many activities.
The SPR Assessment Approach
The SPR (Software Productivity Research) assessment interview approach is a group activity for each project analyzed. The SPR consultant meets with the project manager and as many as seven technical staff members at the same time. All participants have seen the questionnaires ahead of time.
The "official" assessment questionnaire is filled out by the SPR consultant using group consensus as the source of information. Sometimes there may be debate or polarized opinions among team members, and a full set of responses is recorded. These differences are then resolved.
The SPR assessment approach uses structured interviews based on a set of scripted questions. The questionnaires make use of sets of 100 or more than 250 multiple-choice questions. Some SPR questions overlap the same topics as SEI questions, although the forms are different.
Once a sample of projects has been analyzed using the questionnaires, the answers are input directly into a tool that performs statistical analysis of the results and shows the mean values and standard deviations of all factors.
๐ก Why this matters: Although strengths are very laudable, the real value of a process assessment for most clients is the discovery and objective analysis of areas where the client is lagging and needs improvement. Identifying these areas of weakness is the key to a successful process improvement program.
Six-Stage Improvement Program
Once an assessment is completed, the patterns of strengths and weaknesses observed usually lead to a desire to improve software development processes. Organizations can follow a six-stage improvement program that provides an overall framework for software improvement strategies.
Stage 1: Focus on Management Technologies Stage 1 concentrates on management issues, which are critical to all downstream activities. Software process improvement is a fairly expensive undertaking that can stretch out over three to five years. Unless the management community is at state-of-the-art levels in cost estimating, value analysis, and risk analysis, funding for process improvement may be denied. Management needs training in critical technologies such as planning, sizing, estimating, tracking, measurement, risk analysis, and value analysis. Managers collect ROI (Return on Investment) and data that demonstrate progress.
Stage 2: Focus on Software Processes and Methodologies Stage 2 involves the introduction of specific process technologies, e.g., formal code and design inspections. These improvements require training and cultural adjustments but are not very expensive in terms of capital equipment or major investments. Solid approaches for dealing with requirements, design, development, and quality control are implemented. According to a study, projects that used no formal methods at all lagged in terms of productivity, schedules, defect prevention, and defect removal efficiencies against those projects which used them.
Stage 3: Focus on New Tools and Approaches Stage 3 improvements may involve fairly costly tool suites and capital equipment, such as better workstations. As tools support processes rather than the other way around, process improvements precede changes in tool suites. Organizations use various tools and explore new or advanced technologies. Studies show the difference between leading and lagging companies is the use of quality assurance tools, testing tools, and project management tools.
Stage 4: Focus on Infrastructure and Specialization Stage 4 improvements deal with corporate "infrastructures." An example is the establishment of formal quality assurance departments, formal measurement departments, and testing and maintenance departments. As infrastructure changes are both expensive and complex, they need to be deferred until the value of previous kinds of changes has started to become obvious. This stage addresses issues of organization and specialization, moving toward the establishment of specialized teams for handling testing, maintenance, integration, and configuration control, along with formal quality assurance departments and policies on continuing education.
Stage 5: Focus on Reusability Stage 5 improvements are the most valuable in terms of improving quality and productivity, but also the most challenging. Stage 5 moves toward a full and formal software reusability program. Reuse is a double-edged sword โ if reusable materials are of top quality, reuse has the best positive ROI of any software technology. However, if reusable materials are error prone and filled with defects, the ROI switches from positive to sharply negative.
Successful software reuse demands state-of-the-art quality control, software measurements, and software management training.
Reusable Components include:
- Reusable architecture
- Reusable requirements
- Reusable plans
- Reusable estimates
- Reusable designs and specifications
- Reusable interfaces
- Reusable data
- Reusable screens or screen elements
- Reusable source code
- Reusable user documents
- Reusable test plans
- Reusable test cases
Stage 6: Focus on Industry Leadership Stage 6 involves using software excellence for competitive purposes, such as winning a Baldrige Award and using this award to aid in gaining market share. Characteristics include executives who understand and support software, software project managers and technical staff of exceptional ability, more specialists than average organizations, better measurements, better customer satisfaction and staff morale levels, and powerful tool suites for project management, software development, quality assurance, testing, and geriatric tools for aging software.
The Cost of Process Improvements
Every company needs to create an individualized plan and budget for its improvement strategy. Cost elements include training, consulting fees, capital equipment, software licenses, and improvement in office conditions.
The Timing of Process Improvements
Smaller companies move much more rapidly than larger corporations and government agencies. Process improvement is a multi-year undertaking. When there is polarization of opinion or political opposition, progress can be very slow or nonexistent.
The Value of Process Improvements
The lecture notes indicate this section was not filled in the provided text.
โญ Key Takeaways
Software process assessments serve as diagnostic tools, similar to annual medical checkups, that identify strengths and weaknesses in software development practices โ they are not therapeutic but diagnostic. The ten key process areas and ten personnel-related topics examined during assessments have significant impacts on project schedules, productivity, and quality. The SEI and SPR assessment methodologies use different approaches: SEI uses on-site interviews with five maturity levels of CMMI, while SPR uses group consensus with statistical analysis of multiple-choice questionnaires. The six-stage improvement program sequentially addresses management technologies, software processes, tools, infrastructure, reusability (which has the best ROI but requires state-of-the-art quality control), and industry leadership โ and must be viewed as a multi-year undertaking rather than a quick fix.
๐ง Quick Revision Questions
-
What is the primary purpose of software process assessments, and how does the medical examination analogy explain their role in software improvement?
-
List the ten key process areas examined during software process assessments and explain why these specific areas are commonly assessed across different methodologies.
-
Compare and contrast the SEI assessment approach with the SPR assessment approach in terms of data collection methods, interview structure, and outcome analysis.
-
What are the six stages of the software improvement program? Explain why stage 5 (reusability) is described as both the most valuable and the most challenging.
-
Why is software reusability described as a "double-edged sword," and what conditions are necessary for achieving positive ROI from software reuse?