Introduction to Software Engineering Somnuk Keretho, Assistant Professor Department of Computer Engineering Faculty of Engineering, Kasetsart University Email: sk@nontri.ku.ac.th URL: http://www.cpe.ku.ac.th/~sk Somnuk Keretho/Kasetsart University
Outline of this presentation Scope of Software Engineering Object-Oriented Software Development Software Process Software Life-Cycle Models Object Orientation Software Quality Assessment Reference to Chapter 1/2/3 of Software Engineering with JAVA, S.R. Schach, McGraw-Hill, 1997. Somnuk Keretho/Kasetsart University 2
Scope of Software Engineering Software engineering is a discipline whose aim is the production of fault-free software, that is delivered on time, within budget, and satisfies the user s needs. Somnuk Keretho/Kasetsart University 3
Scope of Software Engineering Historical Aspects: 1967, a NATO group coined the term Software Engineering 1968 NATO Software Engineering Conference concurred that Software production should be an engineering-like activity. Using philosophies and paradigms of established engineering disciplines to solve Software Crisis: that the quality of software was generally unacceptably low and that deadlines and cost limits were not being met. Somnuk Keretho/Kasetsart University 4
Scope of Software Engineering Economic Aspects Software Engineering v.s. Computer Science The computer scientist investigates several ways to produce software, some good and some bad. But the software engineer is interested in only those techniques that make sound economic sense. For example: A coding technique that can execute very efficiently but with higher maintenance cost may not be a good choice. Somnuk Keretho/Kasetsart University 5
Scope of Software Engineering Maintenance Aspects Software Life Cycle / Software Process Requirements Phase Specification (Analysis) Phase Planning Phase Design Phase Implementation Phase Integration Phase Maintenance Phase (highest cost among all these phases) Corrective, Perfective, and Adaptive Maintenance Retirement Somnuk Keretho/Kasetsart University 6
Scope of Software Engineering Maintenance is so important, a major aspect of software engineering consists of techniques, tools, and practices that lead to a reduction in maintenance cost. Maintenance 67% Requirement 2% Specification 4% Planning 1% Design 6% Module Coding 5% Module Testing 7% Integration 8% Maintenance 76% Approximate relative costs of the phases of the software life cycle. Somnuk Keretho/Kasetsart University 7
Scope of Software Engineering Specification and Design Aspects Software professionals are humans, and humans can make error. The fact that so many faults are introduced early in the software life cycle, highlights another important aspects of software engineering, namely, techniques that yield better specifications and designs. For example, reducing specification and design faults by 10% will reduce the overall number of faults by 6-7%. Somnuk Keretho/Kasetsart University 8
Scope of Software Engineering Team Programming Aspects Most software being developed and maintained by a team of software engineers Scope of software engineering must also include techniques for ensuring that teams are properly organized and managed. For example, team programming leads to interface problems among code components and communication problems among team members. Somnuk Keretho/Kasetsart University 9
Scope of Software Engineering Several techniques have been suggested to help solve the software crisis. ~1975-1985: Structured Paradigm Structured Systems Analysis, Composite/Structured Design, Structured Programming, Structured Testing Lead to major improvements for software industry. But only good for small programs (say, 5,000-50,000 lines of codes) Not scale well with today larger programs (say, 500.000-5,000,000 LOC) Not so good in software maintenance aspects, (for instance, because of the separation of action-oriented and data-oriented in structured paradigm). Object-Oriented Paradigm An object is a unified software component that incorporates both data and actions that operate of those data. Somnuk Keretho/Kasetsart University 10
Scope of Software Engineering Structured Paradigm Requirement Phase Specification (Analysis) Phase Planning Phase Design Phase Implementation Phase Integration Phase Maintenance Phase Retirement Object-Oriented Paradigm Requirement Phase Object-Oriented Analysis Phase Planning Phase Object-Oriented Design Phase Object-Oriented Programming Phase Integration Phase Maintenance Phase Retirement Comparison of life cycles of structures paradigm and object-oriented paradigm. Somnuk Keretho/Kasetsart University 11
Object-Oriented Software Development Three key words. Software Development Object Orientation Let us look at each in turn Somnuk Keretho/Kasetsart University 12
Software Programs Documentation during the development of programs (e.g. specification) Primary aids for running the programs (e.g. user manuals) Secondary aids for running the programs (e.g. key boards overlays) Software is not just programs! Somnuk Keretho/Kasetsart University 13
Software Life Cycle Software is like humans. It has a life cycle. Software in a system is conceptualized first. It becomes obsolescent at the end. The period in between is called the software life cycle. Somnuk Keretho/Kasetsart University 14
Software Life Cycle Models Build-and-Fix Model Waterfall Model Rapid prototyping model Incremental Model Spiral Model Concurrent Development Model Formal Methods Model For the first four items, please refer to Chapter 3 of Software Engineering with JAVA, S.R. Schach, McGraw-Hill, 1997. Somnuk Keretho/Kasetsart University 15
Built-and-Fix Model Unfortunately, many s/w products are developed with built-and-fix model. Without specification or any attempt in design, just build a product, and reworked as many times needed to satisfy the customer. Unsatisfactory for any size of s/w development, we better specify the various phases of software process. Somnuk Keretho/Kasetsart University 16
Why use a life cycle model? Life cycle model breaks down the development process into phases or stages. This is because software development is complex. Breaking down the development process makes it easier to manage. Each phase can be performed in various ways. Somnuk Keretho/Kasetsart University 17
Waterfall Model Requirement Verify Changed Requirements verify Specification Verify Planning Verify Design Verify Implementation Testing Integration Testing Development Maintenance Operation Mode Retirement Somnuk Keretho/Kasetsart University 18
Rapid Prototyping Model A rapid prototype is a working model that is functionally equivalent to a subset of the product (internal structure is not concerned yet). The sole use of rapid prototyping is to determine what the client s real needs are, construct the rapid prototype as rapidly as possible to speed up the s/w development process. Somnuk Keretho/Kasetsart University 19
Rapid Prototyping Model Rapid Prototype Verify Changed Requirements verify Specification Verify Planning Verify Design Verify Implementation Testing Integration Testing Development Maintenance Operation Mode Retirement Somnuk Keretho/Kasetsart University 20
Incremental Model The s/w product is designed, implemented, integrated, and tested as a series of incremental builds, where a build consists of code pieces from various modules interacting to provide a specific functional capability. It is sometimes necessary to re-specify, re-design, re-code, or at worst, throw away what has already been completed and start again. Somnuk Keretho/Kasetsart University 21
Incremental Model Requirement Verify Specification Verify Planning Verify Architectural Design Verify For each build: Perform detailed design, implementation, and integration. Test. Deliver to client. Development Maintenance Operation Mode Retirement Somnuk Keretho/Kasetsart University 22
Spiral Model The idea of minimizing risk via the use of prototypes and other means is the concept underlying the spiral model. A simplified spiral model is as a waterfall model with each phase preceded by risk analysis. Before commencing each phase, an attempt is made to control (resolve) the risks. If it is impossible to resolve all the significant risks at a stage, then the project is immediately terminated. Somnuk Keretho/Kasetsart University 23
Full Spiral Model [Boehm, IEEE 1998] Review Determine objectives, alternatives, constraints Commitment Partition Plan next phase Requirement plan life-cycle plan Cumulative cost Development Plan Integration and Test Plan Progress through steps Risk Analysis Concept of Operation Requirement Validation Design validation and verification Implementation Risk Analysis Risk Analysis Prototype 1 Prototype 2 Prototype 3 Software Requirements Acceptance Test Evaluate alternatives, identify, resolve risks Risk Analysis Operational Prototype Simulations, models, benchmarks Software Product Design Integration Test Unit Test Detailed Design Code Develop, verify next-level product Somnuk Keretho/Kasetsart University 24
Software Development Software is developed using a life cycle model. Just a life cycle model is insufficient for development. We need: A broad philosophy A set of tools which support the philosophy. A language which supports the philosophy. Somnuk Keretho/Kasetsart University 25
Software Development Paradigm A paradigm provides a general approach to work during each phase of the life cycle model. A paradigm is a broad philosophy. A paradigm is not a specific model. Somnuk Keretho/Kasetsart University 26
Some Software Development Paradigms Functional Composition Logic Programming Structured Development Object Orientation Somnuk Keretho/Kasetsart University 27
Functional Development A problem is expressed in termed of a set of mathematical functions. e.g. Double(x) = Add(x, x). An algorithm is not specified. Language such as Miranda, Gofer, Haskell support this paradigm. Poor execution speed. Somnuk Keretho/Kasetsart University 28
Logic Programming Consists of a problem description only. e.g. Factorial(0) = 1. Factorial(N) = N x Factorial(N -1). Doesn t describe how to solve the problem. Languages Prolog & Lisp support this paradigm. Somnuk Keretho/Kasetsart University 29
Structured Development Also called SASD, SADT & Functional Decomposition. Breaks the system into processes & decomposes them. Languages C, Fortran, Pascal, Cobol, Basic and a lot more support this paradigm. By far the most popular paradigm. Somnuk Keretho/Kasetsart University 30
Object Orientation Most recent paradigm. Treats a problem as a collection of objects. Becoming very popular now. More and more languages support this paradigm now. Somnuk Keretho/Kasetsart University 31
Tools for OO Rambaugh (OMT) Coad-Yourdon Booch UML Somnuk Keretho/Kasetsart University 32
Languages for OO C++ Smalltalk Eiffel Object C Object Pascal Java Somnuk Keretho/Kasetsart University 33
Software Quality Assessment CMM by SEI ISO 9000 Somnuk Keretho/Kasetsart University 34