Object-Oriented Software Design and Software Processes

Transcription

1 Object-Oriented Software Design (COMP 304) Object-Oriented Software Design and Software Processes Hans Vangheluwe Modelling, Simulation and Design Lab (MSDL) School of Computer Science, McGill University, Montréal, Canada Hans Vangheluwe 1/22

2 Overview 1. (Software) Process: definition 2. Various Software Processes 3. The Process Influences Productivity: Dynamic Process Modelling using Forrester System Dynamics Hans Vangheluwe 2/22

3 Process: A Queueing System Arrival Queue Physical View Cashier Departure Departure Arrival [IAT distribution] Queue Cashier [ST distribution] Abstract View Hans Vangheluwe hv@cs.mcgill.ca 3/22

4 Event/Activity/Process Cust2 Process Cust2 Activity queue Cust2 Activity pay cashier Cust1 Process Cust1 Activity pay cashier Cust1 Arrival Cust1 Start pay cashier Cust1 End pay cashier Cust1 Leave t Cust2 Arrival Cust2 Start Queueing Cust2 End Queueing Cust2 Start pay cashier Cust2 End pay cashier Cust2 Leave Event Hans Vangheluwe hv@cs.mcgill.ca 4/22

5 Software Processes The Software Engineering process is the total set of Software Engineering activities needed to transform requirements into software. Watts S. Humphrey. Software Engineering Institute, CMU. Hans Vangheluwe 5/22

6 Software Processes (see notes) Waterfall (Royce) V Model (German Ministry of Defense) Prototyping Operational Specification Transformational (automated software synthesis) Phased Development: Increment and Iteration Spiral Model (Boehm) The Rational Unified Process (RUP) Extreme Programming (XP) Hans Vangheluwe hv@cs.mcgill.ca 6/22

7 The Rational Unified Process (RUP): Activity Workload as Function of Time Hans Vangheluwe 7/22

8 The Rational Unified Process (RUP): Observations 1. Waterfall-like sequence of Requirements, Design, Implementation, Testing. 2. Not pure waterfall: Phased Development (iterative) Overlap (concurrency) between activities 3. Testing: Regression (test not only newly developed, but also previously developed code) Testing starts before design and coding (Extreme Programming) Hans Vangheluwe 8/22

9 RUP: Phased Development Hans Vangheluwe 9/22

10 Extreme Programming (XP) Hans Vangheluwe 10/22

11 Extreme Programming (XP) highlights User Stories are written by the customers as things that the system needs to do for them. They drive the creation of acceptance tests. The project is divided into Iterations. Hans Vangheluwe 11/22

12 Extreme Programming (XP) highlights Use Class, Responsibilities, and Collaboration (CRC) Cards to design the system. Hans Vangheluwe 12/22

13 Extreme Programming (XP) highlights Code the Unit Test first. All code must have Unit Tests; All code must pass all unit tests before it can be released. Refactor whenever and wherever possible. Hans Vangheluwe 13/22

14 Extreme Programming (XP) highlights Pair Programming jriley/pairall.html Hans Vangheluwe 14/22

15 The Process influences Productivity Adding manpower to a late software project makes it later. Fred Brooks. The Mythical Man-Month. Hans Vangheluwe hv@cs.mcgill.ca 15/22

16 Why Brooks Law? Team Size. Model in Forrester System Dynamics using Vensim PLE ( development rate = nominal_productivity* (1-C_overhead*(N*(N-1)))*N Hans Vangheluwe hv@cs.mcgill.ca 16/22

17 Team Size N = 5 Hans Vangheluwe hv@cs.mcgill.ca 17/22

18 Team Size N = Optimal Team Size between 7 and 8 Hans Vangheluwe hv@cs.mcgill.ca 18/22

19 The Effect of Adding New Personnel (FSD model) development rate = nominal_productivity* (1-C_overhead*(N*(N-1)))* (1.2*num_exp_working + 0.8*num_new) Hans Vangheluwe hv@cs.mcgill.ca 19/22

20 5 New Programmers after 100 days Hans Vangheluwe 20/22

21 5 New Programmers after 100 days Hans Vangheluwe 21/22

22 New Programmers after 100 days Hans Vangheluwe hv@cs.mcgill.ca 22/22