CSCI 6231: Software Engineering Instructor: Morris Lancaster Web Page http://www.seas.gwu.edu/~mlancast/csci6231 CSCI 6231 Lecture 0 2 Instructor Morris Lancaster – (703) 400 6408 – lancaster_morris@bah.com – Office hours by appointment CSCI 6231 Lecture 0 3 SWE Courses • Domain Specific – Aeronautical/Aerospace • Level of Detail/Process Oriented – Overview course from a civil/environmental eng. Dept – Elements of Software Construction from a CS department CSCI 6231 Lecture 0 4 Openly Available Resources • Software Engineering Courses/Course Notes – MIT Open Courseware – Open Software Engineering – Free Open Courseware Page • Open Source Tools – At the Development Level Another – At the Project Management Level Another • Professional Societies – Association for Computing Machinery SigSoft – IEEE CSCI 6231 Lecture 0 5 Software Engineering “the application of engineering to software” – IEEE Std 610.12 -1990, Standard Glossary of Software Engineering Terminology (ANSI) “the multi-person construction of multi-version software” - Parnas CSCI 6231 Lecture 0 6 Course Approach Present the background body of knowledge found in the text Use a team project to provide some hands on experience for a development effort (the teams will define, estimate, plan, provide management reviews, develop, test, implement and present their projects) CSCI 6231 Lecture 0 7 Programming versus Software Engineering Programming – a personal activity Structured programming did not solve our problems Software Engineering – team activity Software Engineers emerged (honest brokers, heroes, and snake oil salesmen) CSCI 6231 Lecture 0 8 Role of Software Engineering In overall system design Software is a component of larger systems Software engineers now participate in the design of the whole system (some arrogance among systems engineers still remains, which often times pushes them on to the next project before the roof caves in) CSCI 6231 Lecture 0 9 Software Engineering Choices Trade-off Hardware, software, or elbow wear Compromise Good, fast, cheap (pick 2) Performance, schedule, cost CSCI 6231 Lecture 0 10 Software Engineering Environment Risk Cost Schedule Performance We hope to address some principles that will withstand the changes in how we develop large software products CSCI 6231 Lecture 0 11 Software Engineering History In the beginning was the machine language programmer who addressed well understood problems that had to be solved quickly Computers became cheaper and high level languages emerged - programming became a profession - user and programmer became separate - user wrote specification and programmer implemented - still very few large projects - reuse through subroutines CSCI 6231 Lecture 0 12 Software Engineering History 1960’s - MIT (MULTICS Project) link link - IBM OS 360 (Over 1 staff century of labor) -- Fred Brooks – Mythical Man Month link - Scaling of development efforts was difficult - folklore - Structured Programming – Harlan Mills - The problems being solved were not well understood (lots of basic research in operating systems conducted) - Basic concepts emerged, reuse through libraries, modularity, pathological coupling (later used in structured design) CSCI 6231 Lecture 0 13 Software Engineering History 1970’s - Even cheaper computers micro mini - Range of problems attacked grew - Many failures - Emergence of structured design -- cohesion -- coupling - Team efforts addressed - Still much communication by folklore CSCI 6231 Lecture 0 14 Software Engineering History 1980’s - Even cheaper computers and larger number of programming languages - DoD attempts to solve manpower problems by mandating Ada - In the programming genre encapsulation and abstract data types emerged - Serious metrics and whole life cycle methodologies emerged CSCI 6231 Lecture 0 15 Software Engineering History 1990’s - Even cheaper computers, industry settles on C/C++, Academia uses C++ and moves toward Java, Ada all but dies - Java language was developed - Industry products developed that incorporated business logic (the reuse concept that began with subroutines, matured with abstract data types and object oriented development, was now applied to whole areas of business) - Whole life cycle methods continued, new development concepts emerge to break the long time frame of the waterfall model – rapid prototyping, agile development - Components emerge from middleware CSCI 6231 Lecture 0 16 Engineering - The solution to an engineering construction or design problem can either be visualized, or is the result of a selection from a possible set of alternatives that can be “constructed” directly by piecing together components with known characteristics and capabilities. (We can visualize a house, draw a scale model, pick the appropriate components to meet the underlying requirements at each point, and produce a result. CSCI 6231 Lecture 0 17 The Software Life Cycle - Requirements Analysis and Specification (expressed in some form A) - System Design and Specification (expressed in form A and B) - Coding and Testing (results in a program and the necessary artifacts to document the program, expressed in form C) - Integration and system testing (parts are put together and tested, test results communicated in form D) - Delivery and Maintenance (system is operational, with changes to remove defects and meet new requirements) (THE WATERFALL MODEL) How do we connect these representations so that the transitions can be verified? CSCI 6231 Lecture 0 18 Relationship of SWE to Other Areas of Computer Science Programming Languages - modularity, encapsulation, reuse Operating Systems - virtual machines, portable OS Databases - most computation is processing information Artificial Intelligence - large, complex, and not fully understood systems were developed Theoretical Models - from 210, remember the logical and algebraic paradigms, CSCI 6231 Lecture 0 19 Relationship of SWE to Other Disciplines Management Science As much of an oxymoron as software engineering? Not really, the difficulty is in understanding the degree of precision with which we can work. Systems Engineering Leave the difficult problems for the software? CSCI 6231 Lecture 0 20 Chapter 1 Some Key Points of Chapter 1 - Two kinds of challenges in software development, essential and accidental -- accidental problems can be remedied -- essential problems are more of an issue - Complexity is our enemy, it increases risk -- complexity of the problem -- complexity due to size of project CSCI 6231 Lecture 0 21