Metric Based Software Project Management

1. Metric Based Software Project Management Ashkan Sami, Ph. D. Department of CSE and IT Shiraz University asami@ieee.org sami@shirazu.ac.ir

2. Software project management • Concerned with activities involved in ensuring that software is delivered on time and on schedule and in accordance with the requirements of the organisations developing and procuring the software. • Project management is needed because software development is always subject to budget and schedule constraints that are set by the organisation developing the software.

3. Software management distinctions • The product is intangible. • The product is uniquely flexible. • Software engineering is not recognized as an engineering discipline with the sane status as mechanical, electrical engineering, etc. • The software development process is not standardised. • Many software projects are 'one-off' projects.

4. Management activities • Proposal writing. • Project planning and scheduling. • Project costing. • Project monitoring and reviews. • Personnel selection and evaluation. • Report writing and presentations.

5. Software Project Management • Start of the project • Build the Macro Model • Build the Detailed Project Plan • Build the Project Team • How to Run the Project on a Day-to-Day Basis • Monitoring and Controlling the Project • Management style • Performance Management • Performance Analysis • Interfaces • Systems • Successfully Shutting Down the Project

6. Monitoring and Controlling the Project To monitor and control effectively: • your management style; • management of project performance; • analysis of project performance; • how you will manage the project interfaces; • what systems you are going to use.

7. MANAGEMENT STYLE • Three basic styles of management: • the dictatorial style, • the consensus style and • the laissez-faire style. • A good project manager should be able to recognize when a particular style is appropriate. • You should be able to adopt that style when it is required.

10. PERFORMANCE MANAGEMENT • Performance and its control is one of the areas on which you should spend a significant amount of time. • Sadly, measuring how successful a project is can be difficult, particularly the success of an advanced project. • It is difficult to measure success because advanced projects tend to be unique projects. This means that little in the way of previous performance measures exists.

11. Quality Management • Software Quality Standards • ISO • CMM • Quality Planning • Cost Benefit Analysis • Benchmarking • Cost Of Quality • Quality Management Plan • Quality Assurance • Metrics Measurements • Corrective Actions • Preventive Actions • Quality Control • Audits, Analysis • Checklist

12. Software Project Quality Management • The Activities that Determine Quality POLICIES, OBJECTIVES, RESPONSIBILITIES. • It includes the PLANNING, ASSURANCE and CONTROL of the POLICIES, PROCEDURES and PROCESSES • Forecast the End Result Quality during the Development Stages • Special Characteristic of Software: • Increasing Criticality and Importance of Software • Intangibility of software is a source for Unpredicted Errors • Software Errors are accumulating through the project life cycle

13. Software Product Quality Factors • Product Operation: • Correctness (follow specifications), • Reliability (number of errors, fault tolerance, simplicity), • Efficient (optimize resources usage: storage, memory, CPU) • Integrity (access control & audit to processes and data) • Usability (ease of use/training, operation) • Product Revision (changes, enhancements) • Testing (effort required, regression tests) • Flexibility (effort required to modify the functionality) • Maintenance (effort required to locate and fix a bug: modularity, no duplication of code, documentation) • Product Flexibility (Transition) • Portability (change h/w, operating system, compiler, etc) • Interoperability (integrate with other systems, flexible data and interface structures) • Reusability (can be used as part of another application as a whole or parts – platform free, modular design)

14. Software Quality Standards • ISO 9001 (International Organization for Standardization) • Defines the needs from a monitor & control system to check quality • Certificationof the Design, Development, Production, Installation an Service Processes • Describes the fundamental features of Quality Management System (QMS) • Useful in selecting a sub-contractor with “best practices” quality processes. • Deals with quality of the development process, not with the quality of the product

15. ISO 9001 Principles • Determine the NEEDS and Expectations of the customer • Establish Quality Policy and imply the actual quality objectives • Design the project activities to include the quality objectives • Assign responsible parties to all the quality objectives • Allocate enough and knowledgeable resources • Implement methods to measure the quality objectives in each process • Collect & Analyze the Measurements and Identify discrepancies • Define action items to eliminate the cause of the discrepancies • Documentation (follows quality manual) of the actual (updated) operation of an activity that includes objectives, plans, procedures and records • The QMS is well managed and knowledgeable resources are assigned to the Quality Management Process. • Demonstrate that the Production Process is well defined, designed, recorded, communicated and measured. • Sub-Contractors and Purchasing are managed in the same manners.

16. CMM – Capability Maturity Model • Software Development Methods and Tools which are Likely to produce Quality Software • Software Companies are assigned a level ( 1 to 5 ) of Process Maturity that indicates the quality of their software production practices • Level 1: Initial Level • Default level, no defined quality process throughout the organization. Some projects may adopt some measures. • Level 2: Repeatable Level • Basic Project Management in place, not in the activities level • Level 3: Defined Level • Project Management Plan is well defined at all levels • Level 4: Managed Level • Products & Processes are Measured and Controlled • Level 5: Optimizing Level • Process Improvement is introduced based on documented data gathered from previous projects, processes.

17. 2 Managed CMMI Levels Level Process Areas Improving, addressing common causes of variation Causal Analysis and Resolution Organizational Innovation and Deployment 5 Optimizing Quantitatively managed, eliminating special causes of variation Quantitative Project Management Organizational Process Performance 4 Quantitatively Managed Requirements Development Technical Solution Product Integration Verification Validation Organizational Process Focus Organizational Process Definition Organizational Training Risk Management Integrated Project Management (for IPPD*) Integrated Teaming* Integrated Supplier Management** Decision Analysis and Resolution Organizational Environment for Integration* Work proactively managed, organizational standard processes 3 Defined Work planned and tracked (reactively managed) Requirements Management Project Planning Project Monitoring and Control Supplier Agreement Management Measurement and Analysis Process and Product Quality Assurance Configuration Management Work performed, but in an ad hoc fashion 1 Performed

18. CMMI Staged Representation - 5 Maturity Levels Process performance continually improved through incremental and innovative technological improvements. Level 5 Optimizing Level 4 Quantitatively Managed Processes are controlled using statistical and other quantitative techniques. Process Maturity Level 3 Processes are well characterized and understood. Processes, standards, procedures, tools, etc. are defined at the organizational (Organization X )level. Proactive. Defined Level 2 Managed Processes are planned, documented, performed, monitored, and controlled at the project level. Often reactive. Level 1 Initial Processes are unpredictable, poorly controlled, reactive.

19. Maturity Level 1 Initial • Maturity Level 1 deals with performedprocesses. • Processes are unpredictable, poorly controlled, reactive. • The process performance may not be stable and may not meet specific objectives such as quality, cost, and schedule, but useful work can be done.

20. Maturity Level 2Managed at the Project Level • Maturity Level 2 deals with managed processes. • A managed process is a performed process that is also: • Planned and executed in accordance with policy • Employs skilled people • Adequate resources are available • Controlled outputs are produced • Stakeholders are involved • The process is reviewed and evaluated for adherence to requirements • Processes are planned, documented, performed, monitored, and controlled at the project level. Often reactive. • The managed process comes closer to achieving the specific objectives such as quality, cost, and schedule.

21. Maturity Level 3Defined at the Organization Level • Maturity Level 3 deals with defined processes. • A defined process is a managed process that: • Well defined, understood, deployed and executed across the entire organization. Proactive. • Processes, standards, procedures, tools, etc. are defined at the organizational (Organization X ) level. Project or local tailoring is allowed, however it must be based on the organization’s set of standard processes and defined per the organization’s tailoring guidelines. • Major portions of the organization cannot “opt out.”

22. Behaviors at the Five Levels Optimizing Defined Over reliance on experience of good people – when they go, the process Managed goes. “Heroics.” Maturity Level Process Characteristics Behaviors Focus on "fire prevention"; Focus is on continuous quantitative improvement improvement anticipated and desired, and impacts assessed. Greater sense of teamwork and inter- Process is measured and controlled Quantitatively Managed dependencies Reliance on defined process. People understand, support and follow the process. Process is characterized for the organization and is proactive Process is characterized for projects and is often reactive Focus on "fire fighting"; Process is unpredictable, poorly controlled, and reactive Initial effectiveness low – frustration high.

23. CMMI Components • Within each of the 5 Maturity Levels, there are basic functions that need to be performed – these are called Process Areas (PAs). • For Maturity Level 2 there are 7 Process Areas that must be completely satisfied. • For Maturity Level 3 there are 11 Process Areas that must be completely satisfied. • Given the interactions and overlap, it becomes more efficient to work the Maturity Level 2 and 3 issues concurrently. • Within each PA there are Goals to be achieved and within each Goal there are Practices, work products, etc. to be followed that will support each of the Goals.

24. CMMI Process Areas

25. State of the Software Industry in 1995 • Maturity Level Frequency 1. Initial 70% 2. Repeatable 15% 3. Defined < 10% 4. Managed < 5% 5. Optimizing < 1% Source: Royce, Project Management, P. 364

26. با تشکر از توجه شما تقدیم به روح پر فتوح یکی از اولین فارغ التحصیلان و اساتید دانشگاه علوم پزشکی دکتر منصور سامی

27. Identify the Project Quality Standards Determine how to Satisfy the Standards Inputs: Enterprise Factors (law, regulations, standards) Organizational Factors ( QA policies) Scope Statement (objectives, thresholds, acceptance criteria, performance criteria, etc.) Project Management Plan Tools & Techniques Cost Benefit Analysis Cost Of Quality: Expenses associated with the Project Quality Management Benefit: Less Rework, higher productivity, lower cost, higher stakeholder satisfaction Benchmarking Compare the project process practices to other projects to get ideas for improvements Others: Brainstorming, Flowcharts Quality Planning

28. Identify the Project Quality Standards Determine how to Satisfy the Standards Inputs: Enterprise Factors (law, regulations, standards) Organizational Factors ( QA policies) Scope Statement (objectives, thresholds, acceptance criteria, performance criteria, etc.) Project Management Plan Tools & Techniques Cost Benefit Analysis Cost Of Quality: Expenses associated with the Project Quality Management Benefit: Less Rework, higher productivity, lower cost, higher stakeholder satisfaction Benchmarking Compare the project process practices to other projects to get ideas for improvements Others: Brainstorming, Flowcharts Quality Planning

29. Quality Planning - Output • Quality Management Plan • How the project management will implement the quality policy • Quality Control Plan • Quality Assurance Plan • Continuous Improvement Plan • Quality Metrics • Detailed Plan of what will be measured, actual values • Quality Checklist • Structured tool. Process, Activity based. Templates can be available from prior projects or QA organizations. • Quality Baseline • The quality OBJECTIVES of the project • The basis for measuring and reporting • Process Improvement Plan • Identification of WASTE and NON-VALUE ADDED activities • Set TARGETS for Improved performance

30. Quality Assurance - QA • A QA Department in the Organization • Observe that the project will employ all processes needed to meet Quality Targets • QA support & Knowledge Base • Provides Umbrella for Continuous Process Improvement • Recommend Corrective Actions and Updates to the Project Plan • Update Organizational Process Assets

31. Software Metrics“You Can Not Control What You Can Not Measure” • Retain Records (historical data) – Used to evaluate future projects (sizing,schedule, cost, effort) • Problem Reporting – Supply Information about problem fixing (cycle, time, amount, frequency, etc.) • Quality Assurance Plan – Define Quality Thresholds (What is the definition of “Low System Complexity”) • Corporate Quality Goals – Measure progress toward the Goals • Types of Software Metrics • Directly Observed: Number of Test Cases, Project Cost • Predicted: Project Baseline Cost, Estimate To Complete • Calculated: Productivity=KLOC/Staff-Months

32. Useful Metrics • Precise or Defined with Clear Tolerance, • Well Defined Methods of Data Collection • Helps to Measure the Organization/Project Quality Goals • Simple & Easy to understand • Inexpensive to Use • Robust • Consistent and Used Over Time • Easy To Collect • Easy to Access (online)

33. Software Metrics – Examples of Direct Observations • Control Chart - Upper/Lower Control Limit ( Usually the Corporate or Project Quality Targets) • Actual Measures per Module/Activity VS Corporate Goals. The Average Project Measurement implies the overall project compliance. Deviation from the control limits indicate problem in a specific Module/Activity • Typical Metric: Design Review Hours Invested per Module. • Total Hours to investigate and resolve problem per predefined problem types (interface, documentation, input data, computational, etc.) • Defects Discovered and Defect Fixed Over Time • Mean Age of: Open Problems at the end of the Month Closed Problems at Month End

34. Continuous Process Improvement • Main Goals: • Continuous Improvement in Resource Planning (increase the probability to deliver accurate estimates) • Continuous Reduction in Waste/Non-Productive Project Activities. • Continuous Improvement in Other Corporate Quality Goals • Values in Software Activities • Value Added: Design, Code, Test, Install, Etc. • Non Value Added but Essential: Acceptance Testing, Setup, Training, Etc. • Non Value Added and Nonessential: Retest, Rework, Management, Installing Software Tools, Delays, Etc. • Value Metrics: Value/Total, Non-Value Essential/Non-essential, Value+Essential/Total

35. Quality Control - QC • Monitor SPECIFIC project results to determine its compliance with the Quality Standards • Identify ways to eliminate CAUSES of unsatisfactory results • Continuous Process • Uses Sampling and Probability analysis • Analysis Types • Prevention & Inspection • Sampling (check conformance (Yes/No) to standards, goals, Etc) • Variables Sampling – Level Of Conformance with Standards/Goals • Special Causes/Unusual Events • Tolerance (results fall within/outside the control limits)

36. QC – Tools & Techniques • Cause & Effect Diagram • Per Major Defect (E.G: Users Interface Issues) • Cause: Guidelines not Followed • Reasons: Did not read the guidelines(no time), Can not find the guidelines (No central Location) • Cause: Lack of users feedback • Reasons: Lack of users resources, Prototype not clear enough, No process to get early feedbacks • Control Charts • Flowcharting (activities, decision points, sequencing) – Can point potential quality issues • Histogram – Number of Defects per Problem Type • Pareto Chart – 80/20 rule. Order Defects by Number of Defect and focus on those that causesthe majority

37. QC – Tools & Techniques (Cont.) • Run Charts/Scatter Diagram (indicate trends over time) Examples: • Failure rate (defect found per test hour) over Cumulative test time • New Defects/Resolved defect per week over weeks • Statistical Sampling • Inspection • Defect Repair Reviews