1 / 39

CS 5150 Software Engineering

CS 5150 Software Engineering. Lecture 7 Project Management. Administration. Tests Tests that are not collected in class will be available for pick up from Corinne Russell at 301 College Avenue. Re-grade requests • To request a re-grade, contact me after class or come to office hours.

hendry
Download Presentation

CS 5150 Software Engineering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. CS 5150 Software Engineering Lecture 7 Project Management

  2. Administration Tests Tests that are not collected in class will be available for pick up from Corinne Russell at 301 College Avenue. Re-grade requests • To request a re-grade, contact me after class or come to office hours. • No request will be considered more than one week after the tests are returned. • Grading mistakes will be corrected. Points will not be changed when the grade is determined by the grader's judgment.

  3. Test 1, Question 2 Academic researchers often mount the journal papers that they write on departmental web servers. The goal of the Collection System is to collect all the papers for a large university and load them into a central repository. This will be a record of research at the university, to be preserved for the long-term. The system will look at each departmental server, identify research papers, index them, and copy them to the central repository. (continued on next slide)

  4. Test 1, Question 2 (continued) Here are some of the challenges that are expected: (a) As far as possible, the identification of research papers will be automated. The algorithm for identifying papers is not perfect and errors can be expected. (b) The system will attempt to identify the author of each paper automatically, and send an email message requesting permission to load the paper into the repository. (c) The algorithm for indexing the papers will attempt to extract information such as author, title, and date. (d) Because of errors in (a), (b), and (c), each paper will need human review before loading into the repository. It is important to provide a user interface that minimizes human effort required.

  5. Test 1, Question 2 (continued) You are asked to recommend a software development process for the Collection System. (a) i Is the modified waterfall model a good choice for developing this system? No ii Give two reasons for your answer. (-) The agorithms for automatic processing will need to be tested with real data. This will require iteration. (-) The user interface will need testing with users. This will require iteration. Note that, although the modified waterfall model expects changes during development, it aims at minimal changes and does not support full iteration.

  6. Test 1, Question 2 (continued) (b) i Is iterative refinement leading to a single release of the system a good choice for developing this system? Yes ii Give two reasons for your answer. (+) Iterative refinement allows testing and revision of the user interface to minimize the human effort involved in review. (+) Iterative refinement allows testing and tuning of the automatic algorithms. Note 1. Thorough testing with real data and staff reviewers is essential. Note 2. A single release does not preclude later maintenance.

  7. Test 1, Question 2 (continued) ii Is incremental releases of small increments of software a good choice for developing this system? Possible, but not best Discussion.Can the development be divided into short sprints each involving a team working through a full software cycle including planning, requirements analysis, design, coding, testing, and acceptance testing? Answer. It is possible. Here are some possible sprints: 1. Automatic collection of papers in PDF from engineering departments 2. Human review of papers collected in 1. 3. Automatic generation of emails. etc., etc.

  8. Test 1, Question 2 (continued) (c) ii Give two reasons for your answer. (+) Possible to divide the central system into increments that can be tested and accepted by the client separately (1 point) (-) The details of the review process, with its user interface, will depend on the accuracy of the automated algorithms. This will require iteration. Note. Each increment must be fully tested and accepted before it is considered complete.

  9. Project Management: OS 360 The operating system for the IBM 360 was two years late. Question: How does a project get two years behind schedule? Answer: One day at a time! Fred Brooks Jr., The Mythical Man Month, 1972

  10. Key Personnel: The Mythical Man Month Adding more people may actually slow down a project Adding more people adds communications complexity • Some activities need a single mind. • Sometimes, the elapsed time for an activity can not be shortened. In computing, not all people are equal • The best are at least 5 times more productive. • Some tasks are too difficult for everybody.

  11. The Aim of Project Management To complete a project: • On time • On budget • With required functionality • To the satisfaction of the client • Without exhausting the team To provide visibility about the progress of a project

  12. The Challenge of Project Management Clients wish to know: What will the system do? When will it be delivered? What will it cost? Examples: • If the system is a product, marketing and development must be combined (e.g., Microsoft Office) • If the system has to work with other systems, developments must be coordinated (e.g., embedded systems in an automobile) (continued on next slide)

  13. The Challenge of Project Management (continued) BUT: Every software system is different Estimating time and effort is full of errors, even when the system is well specified Most systems are not well specified, or the specifications change during development

  14. Estimating: Analysis Administrative computing department at Dartmouth used activity graphs for the implementation phase of major projects (plan developed after project was well-understood). Experience: Elapsed time to complete projects was consistently 30% to 40% longer than predicted by model. Analysis: • Some tasks not anticipated (incomplete understanding) • Some tasks had to be redone (change of requirements, technical changes) • Key personnel were on many activities (schedule conflicts) • Non-billable hours

  15. Standard Approach to Project Management • The scope of the project is defined early in the process. • The development is divided into tasks and milestones. • Estimates are made of the time and resources needed for each task. • The estimates are combined to create a schedule and a plan. • Progress is continually reviewed against the plan, perhaps weekly. • The plan is modified by changes to scope, time, resources, etc.

  16. Agile Approach to Project Management • Planning is divided into high level release forecasting and low level detailed planning. • Release planning is a best guess, high level view of what can be achieved it a sequence of time-boxes. • For each time-box, the team plans what it can achieve. • Release plans are continually modified, perhaps daily. • Clients and developers take joint control of the release plans and choice of sprints.

  17. Common Aspects of all Approaches Planning • Outline schedule during feasibility study (needed for CS 5150) • Fuller schedule for each part of a project (e.g., each iteration or sprint) Progress tracking • Regular comparison of progress against plan • Regular modification of the plan • Changes of scope, etc. made jointly by client and developers Final analysis • Analysis of project for improvements during next project (see Dartmouth example)

  18. Estimating the Time for an Activity • With experienced staff, estimating the actual time to carry out a • single task is usually fairly accurate, but ... • The little bits and pieces are underestimated • The time from almost "done" to completely "done" is much longer than anticipated. (There's just one thing to tidy up. I need to put the comments into better shape. I really should get rid of that patch.) • The distractions are not planned for. (My system crashed and I decided to upgrade the software. My child's school was closed because of snow. I spent the day showing visitors around.) • • Some things have to be done twice.

  19. Team-based Estimating • The team often has the best understanding of what it can achieve in a single time-box or sprint. • The team commits to the outcome of a sprint. • The team must have an internal schedule to allocate tasks within a sprint. • Since different teams work at different speeds it is common to estimate effort to achieve a specific goal in a numeric scale, not as time. A CS 5150 project can be thought of as a single sprint.

  20. Start-up Time On a big project, the start-up time is typically three to six months: • Personnel have to complete previous projects (fatigue) or be recruited. • Hardware and software has to be acquired and installed. • Staff have to learn new domain areas and software (slow while learning). • Clients may not be ready.

  21. Methods for Task Based Project Planning Critical Path Method, Gantt charts, etc. • Build a work-plan from activity data. • Display work-plan in graphical and tabular form. Project planning software • Maintain a database of activities and related data • Calculate and display schedules • Provide progress reports

  22. Task Based Project Planning: Terminology Task Part of a project that takes place over time (also known as an activity). Event The completion of a group of activities. Dependency An activity that cannot begin until some event is reached Resource Staff time, equipment, or other limited resources required by an activity.

  23. Task Based Project Planning: Terminology Deliverable Work product that is provided to the client (report, presentation, documentation, code, etc.) Milestone Completion of a specified set of activities (e.g., delivery of a deliverable, completion of a process step, iteration, or sprint)

  24. A Simple Gantt Chart Source: Microsoft using Excel

  25. Gantt Charts Used for Small Projects, Single Time-Boxes, and Sprints • Dates run along the top (days, weeks or months). • Each row represents an activity. Activities may be sequential, in parallel or overlapping. • The schedule for an activity is a horizontal bar. The left end marks the planned beginning of the task. The right end marks the expected end date. • The chart is updated by filling in each activity to a length proportional to the work accomplished. • Progress to date can be compared with the plan by drawing a vertical line through the chart at the current date.

  26. A More Complex Gantt Chart Source: SmartDraw

  27. Activity Graph A scheduling technique that emphasizes dependencies An activity (task) A dummy activity (dependency) An event A milestone

  28. Example: Activity Graph for a Distance Learning Course (part of a large project) Script TV Plan TV Write 1 Edit 1 Make TV START Typeset 1 Plan 1 Print Mail Plan 2 Typeset 2 Write 2 Edit 2 Document Software Write Software Plan Software Prototype Software

  29. Scheduling: Background PERT Program Evaluation and Review Technique introduced by the U.S. Navy in 1957 to support the development of its Polaris submarine missile program. PERT/Time Activity graph with three time estimates (shortest, most probable, longest) on each activity to compute schedules. PERT/Cost Added scheduling of resources (e.g., facilities, skilled people, etc.)

  30. Critical Path Method Uses an Activity Graph with single time estimate for each activity. A standard method for managing large construction projects. On big projects, activity graphs with more than 10,000 activities are common.

  31. Critical Path Method The following slides work through the steps in calculating earliest and latest start dates.

  32. Time Estimates for Activities (Weeks) 4 1 3 2 6 3 1 1 12 3 12 3 2 2 8 4 4

  33. Earliest Event Dates and Earliest Start Dates Earliest start date: the earliest date that it is possible to start an activity. Earliest event date: the date that all the activities ending at that node will be completed, assuming that every activity begins at its earliest start date. Earliest project completion date: the date on which the project is completed assuming that every activity begins on its earliest start date.

  34. Earliest Event Dates and Earliest Start Dates 15 24 22 25 17 17 Earliest dates for an event are in red 1 4 1 5 3 2 6 17 12 23 3 20 1 12 23 1 0 15 3 12 3 2 19 2 12 8 4 8 4 4

  35. Latest Event Dates and Latest Start Dates Latest start date: the latest date that it is possible to start an activity and still complete the project by the earliest project completion date. Latest event date: the latest date that all the activities ending at that node must be completed, in order to complete the project by the earliest project completion date.

  36. Latest Event Dates and Latest Start Dates 19 17 0 17 17 Each event must be achieved by the date shown or the final date will not be met. Latest event dates in blue 11 4 1 20 3 2 6 12 15 3 10 1 12 23 24 25 1 3 12 18 3 2 20 2 14 8 4 22 4 13

  37. Slack and Critical Path Slack: the difference between the latest start date and the earliest start date of an activity Slack = (latest start)end - (earliest start)begin - (time estimate) Critical path: a path through the graph where every activity has zero slack If an activity on the critical path is not started on its earliest start date or takes longer than the predicted time to complete, then the project completion date is delayed. Extra detail added October 3, 2011

  38. Slack 1/11 0/0 Slack in purple 10 4 10 1 12/12 15/15 17/17 6 2 25/25 3 0 2 0 3 22/23 3 23/24 0 12 1 0 1 1 1 12 1 12/14 17/17 19/20 2 3 3 2 1 2 5 0 2 8 4 9 0 4/13 17/17 4 9

  39. Critical Path Critical path in orange Every activity on the critical path must begin on the earliest start date. 1/11 12/12 15/15 17/17 25/25 22/23 23/24 0/0 12/14 17/17 19/20 4/13 17/17

More Related