1、Building a Plan,Instructor: Mike ODell,Several of the slides in this module are a modification and amplification of slides prepared by Mr. Tom Rethard for use in a prior Senior Design Class. They were originally for use with A Discipline for Software Engineering (Watts S. Humphrey), sponsored by the
2、 U.S. Department of Defense. Original slides are copyright SEI; modifications are Copyright 2002, T. Rethard. Further modifications by Mike ODell. All Rights Reserved.,CSE 4316,2,Why Plan?,A plan helps you focus on the goal “Begin with the end in mind.1” A plan lets you estimate job completion A pla
3、n helps you track progress A plan gives you milestones that provide a sense of accomplishment along the way A plan helps you identify problems early A plan establishes commitments for the team and each individual on it,1 Stephen R. Covey, The Seven Habits of Highly Effective People,CSE 4316,3,The Pl
4、anning Process Simplified,Plan the work then work the plan Refine, refine, refine,CSE 4316,4,What is a Plan?,An agreement by the team on the cost and schedule for a specified job A structure for organizing the work A framework for obtaining the required resources (people, funds, etc.) A record of wh
5、at was initially assumed and committed Its a CONTRACT!,CSE 4316,5,Components of a Plan,A Lifecycle Planning Model: The Master Plan for the Project Order and criteria for key events Correct model for the job? Work Estimate How big is the job (size and effort)? How long will it take, when will we fini
6、sh? Schedule and Work Breakdown When do we expect to have things done? What are we committing to?,CSE 4316,6,Selecting the Correct Model,Discussion: Case Study 7.1 Ineffective Lifecycle Model Selection Why was the model selected? What went wrong? What was the result? What might have been done differ
7、ently?,CSE 4316,7,Representative Lifecycle Models,Pure Waterfall (the old granddaddy) Code-and-Fix (and plan to fail) Spiral (new age sophistication) Modified Waterfalls (making the best of an old standby) Evolutionary/Rapid Prototyping (design as you go) Staged Models (show as you go) Design to sch
8、edule Hybrids some combination of above,CSE 4316,8,Pure Waterfall,Phased deliverables Document-driven Discontinuous, inflexible phases All planning is done up front Good for: Well-defined, complex projects Quality dominates cost and schedule Not so good for: Projects where details cannot be complete
9、ly specified up front Rapid development projects,CSE 4316,9,Pure Waterfall,Requirements Analysis,Concept & Planning,Architectural Design,Detailed Design,Implementation & Debugging,System Validation,Pass,Fail,Pass,Pass,Pass,Pass,Fail,Fail,Fail,Fail,CSE 4316,10,Build (Code)-and-Fix,In general, dont do
10、 it! Process: Specify it, code-and-fix it, ship it(?) Advantages: Low/no overhead (no planning, documentation, standards, etc.) You can start TODAY! Requires little other than technical expertise Disadvantages: No means of assessing progress, problems Dangerous! for other than tiny projects,CSE 4316
11、,11,Spiral,Risk-oriented, layered approach Process: Break project into mini-projects, each addressing major risks, e.g. Risk of poor specifications Risk of poorly understood architecture Iterate until risks are eliminated Six steps in each iteration Advantages: Time and money can reduce risk Disadva
12、ntages Complex,CSE 4316,12,Spiral,(Diagram from “Spiral Development: Experience, Principles and Refinements”, workshop paper by Barry Boehm),Start,CSE 4316,13,Modified Waterfalls,Waterfall with Overlapping Phases (Sashimi) Like pure waterfall, but phases can overlap,CSE 4316,14,Modified Waterfalls,W
13、aterfall with Subprojects Begin detailed design on subprojects before overall architectural design is complete,Requirements Analysis,Concept & Planning,Architectural Design,System Validation,CSE 4316,15,Modified Waterfalls,Waterfall with Risk Reduction Spiral for requirements and architectural desig
14、n phases,Requirements Analysis,Concept & Planning,Architectural Design,Detailed Design,Implementation & Debugging,System Validation,Generally, only for very large and complex projects,CSE 4316,16,Evolutionary (or Rapid) Prototyping,Especially useful when requirements are changing rapidly, or cannot
15、be committed Process: Design initial prototype of external/prominent aspects Review with customer Iterate and refine until “good enough” Advantages Keeps customer involved in process Low overhead Disadvantages Impossible to project schedule/budget Can evolve to code-and-fix,CSE 4316,17,Staged Delive
16、ry Models,Follow architectural design with staged design, implementation, test and delivery Staged delivery: iterate until done Design-to-schedule: iterate until scheduled time Evolutionary delivery: Iterate with customer feedback until done (Beta test approach),CSE 4316,18,Agile Methodologies,Itera
17、tive and incremental development Adaptive planning based on customer and market changes Plan milestones are flexible and subject to change “rolling wave” progression TimeBox development Staged (potentially shippable increments) Scrum,CSE 4316,19,Hybrid Staged Delivery Model,Requirements Analysis,Con
18、cept & Planning,Architectural Design,Medium High Priority: Detailed design, implement and test,Medium Priority: Detailed design, implement and test,High Priority: Detailed design, implement and test,Medium Low Priority: Detailed design, implement and test,Low Priority: Detailed design, implement and
19、 test,Run out of time and money,Release,Design-to-Schedule with risk reduction (our model, approx.),CSE 4316,20,Choosing the Right Model,Strengths and weaknesses analysis Discussion: Table 7-1 Case Study 7-2: Effective Lifecycle Model Selection Project characteristics Why was the model the right one
20、? What was the outcome?,CSE 4316,21,Tools/Techniques to Help You,PERT and CPM Tools Program (or Project) Evaluation and Review Technique Critical Path Method (from Dupont) Account for task dependencies Generally applies 3 separate estimates for each task (shortest, nominal and longest) to calculate
21、the expected effort Identify longest/critical path(s),CSE 4316,22,Tools/Techniques to Help You (PERT Chart),CSE 4316,23,Tools/Techniques to Help You,CoCoMo (Constructive Cost Model) Estimating tool created by Barry Boehm (Software Engineering Economics, 1981) Based on size, complexity, environment,
22、team composition, language, tools, etc. Method is based on a large study of varying size significant projects.,CSE 4316,24,Work Breakdown Structure,Breaks down the work to be done into specific, product-oriented manageable units Allows development of a detailed plan Basis for project cost and schedu
23、le Enables assignment of responsibility Provides basis for accountability of individuals Defines independent work units minimum interfacing with or dependency on other work units Allows measurement of progress,CSE 4316,25,Work Breakdown Structure,WBS: Building a Bicycle,CSE 4316,26,Milestone Trackin
24、g - GANTT,GANTT Chart Display Using MS Project,See Sample MS Project file on website.,CSE 4316,27,High Quality Plans,Characteristics, as stated in the SEI text Complete Readily accessible, even by the customer Clear Specific Precise Accurate Not in the SEI text, but necessary MEASURABLE,CSE 4316,28,
25、Whats makes a Good Plan?,Complete Product Specifications A clear Statement of Work Size estimate and schedule Schedule for critical Milestones A complete Work Breakdown Structure The Processes/Procedures that you will follow Identification of your Stakeholders,CSE 4316,29,Whats makes a Good Plan?,Fr
26、om the customers perspective: Your commitment to deliver what is specified The quality level of the product A mechanism for participation/cooperation,Integrity and Openness,CSE 4316,30,Product Specifications,Provide the details of what will be done, and how it will be done: Requirements Specificatio
27、n (SRS what) Architecture Design (ADS bridge what to how) Detailed Design Documentation (DDS - how ) These provide the basis for system verification and acceptance,CSE 4316,31,Statement of Work,A narrative description of the work that is to be done: Details of hardware and software components Descri
28、ption of deliverables Estimate of start and stop dates for key phases of process Acceptance criteria,CSE 4316,32,Milestones,Driven by the lifecycle model you use Establishes start and stop dates for all key phases of project Reinforced by your detailed schedules Use PERT/GANTT Provides basis for mea
29、surement of progress Earned Value Provides basis for identifying and estimating risks Specifies critical deliverables ALL milestones have deliverables!,CSE 4316,33,Processes/Procedures,Defines how things will get done Provides the basis for establishing critical milestones and deliverables Establish
30、es entry and exit criteria for critical phases Establishes the standards that will be used Defines the tools that are required to complete the work,CSE 4316,34,Stakeholders,Any person or organization that has a vested interest in the success of you project Your customer or sponsor Your company Your
31、companys owners/stockholders Your management,CSE 4316,35,Plan Documentation and Tracking,System Requirements Specification WHAT you plan to create Project Charter HOW you will go about the process of creating the WHAT Includes RISK MANAGEMENT plan Work Breakdown Structure/GANTT The specific STEPS yo
32、u will take, WHEN things must be done, and WHO will do them MS Project,CSE 4316,36,Characteristics of a Good Requirement,Verifiable: stated in a way that allows for independent verification that the system meets or exceeds the stated requirement. Justifiable: necessary, rather than simply desirable
33、Unambiguous: stated such that multiple interpretations are precluded,CSE 4316,37,Characteristics of a Good Requirement,Consistent: no conflict with any other requirement Modifiable: should be stated in a way that allows for change based on technical/business considerations. Hierarchically Traceable:
34、 should define a single attribute, traceable back to a higher level requirement.,CSE 4316,38,Tips for Successful Requirement Determination,Start by establishing what the team thinks the features/functions of the system should be Brainstorm as team and write everything down Keep a simple list (such a
35、s the requirements worksheet on the website) Meet with your sponsor to review/modify your list and discuss alternatives Add any features/functions that the sponsor believes are required,CSE 4316,39,Tips for Successful Requirement Determination,Consider and add ancillary requirements E.g., performanc
36、e, packaging, look and feel Discuss and add as necessary any “non-functional” requirements E.g., standards that you must adhere to, maintenance and support, safety Discuss and analyze the feasibility of meeting or exceeding each requirement within the budget, time and skills allowed.,CSE 4316,40,Tip
37、s for Successful Requirement Determination,DO NOT collect requirements by attempting to fill out the SRS Guide!List and understand them, THEN write the document,CSE 4316,41,What is a System Requirements Specification (SRS)?,A detailed description of the features and functions of a product, incorpora
38、ting: End-user and sponsor input Developer input Management input Standards and processes Your documented commitment to deliver Your contract with your stakeholders that identifies WHAT you will create,(See SRSs from prior teams on class website.),CSE 4316,42,What is a Project Charter?,A document th
39、at summarizes the project Defines the scope and objectives of the project Delineates organizational relationships Delineates individual authority and responsibility Identifies key risks and plan to handle them Specifies dependencies on other organizations Describes the specific functions to be performed Lays out the master schedule for the project Documents cost and time estimates Document person-loading requirements/schedule Documents management approval of the details of the project,(See Charter template and Charters from prior teams on class website.),
