Lecture Outlines (DRAFT)

Introduction

• Course organization
• Course content: tutorial, engineering experience, research
• Readings: tutorials, case studies, research papers & diatribes
• Phases of class project: rev eng, design, implementation, reflection
• Why design? Productivity, safety, delegation
• Elements: notations, methods, tools
• Approaches to design: component vs. cross-cutting

Overview of CTAS

• Guest lecture, Prof John Hansman

Entity Life Histories

• Alphabets, events and traces
• Descriptions: regexps, grammars, machines
• Roles, parallelism and shared events
• Premature termination
• Essential vs. inessential structure
• State variables

Basic Object Modelling

• Objects, entities and attributes
• Abstraction of program structures: containers, methods, etc
• Components: sets, relations, indexed relations, overloaded relations
• Subset relationships
• Multiplicity markings
• Organization techniques: ternaries, HO relations, lowering
• Mutability of relations
• Static and dynamic classification
• Derived and designated components
• Omitted features from UML, Catalysis, Syntropy, Fusion, etc

Data Abstraction

• Representation independence
• Type-based enforcement
• Representation exposure
• Constructors, mutators and observers
• Abstraction functions and rep invariants
• Beneficent side effects
• Subclassing exposures
• True parametric polymorphism
• Subclassing and subtyping

Design for Change

• Lessons learnt from John Chapin's course last fall

Design Patterns

• Motivation: handbooks, tricks of trade, Alexander
• Why so popular?
• Structure of a pattern; notation
• Observer: event registration protocols, consistency
• Visitor: subclassing of visitors, subclassing of visited, replacers
• Bridge: Grappa as case study
• Factory

Concurrent Design Patterns

• TBD

Problem Frames

• Decomposition criteria
• Polya's frames
• Phenomena: events, states, facts
• Domains, shared phenomena and characteristics
• Elementary frames
• Composite frames
• Synchrony hypothesis

Software Architecture

• Connectors as 'first-class' objects
• Architectural style
• Sample styles: event systems (Smalltalk, Gandalf, Field)
• Dynamic architectures with object models
• Connector specification and analysis with Wright

Declarative Specification

• Abstract state machines and invariants
• Partitioning the transition relation into operations
• Preconditions and guards
• Non-determinism
• Implicit preconditions
• Implementability, preservation of invariants
• Clever examples: minimization, etc
• Views and multiple representations
• Basic data types: sets, relations, sequences
• Refinement relations

Design Methods

• Principles and criteria: from SRS
• Strong and weak methods
• Formal methods aren't 'methods'
• Benevolent difficulty
• Elements of current OO methods: use cases, OMs, MSCs, etc
• History: generations of methodologies
• Structured analysis: DFDs, dictionaries, structure charts
• Simulation-based methods
• New methods: Fusion, Catalysis

Advanced Object Modelling

• Constraints: definitions, invariants, conditions
• Operations and transitions
• Redundancy: assertions and counterexamples
• Scope and automatic analysis
• Domains and types
• Funky modelling: global states, histories, problem domains
• Being really abstract: orders, timestamps

Concurrency Case Studies

• Guest lecture, Prof. John Chapin

Fads and Fallacies in Software Engineering

• Top-down design and stepwise refinement
• Weyuker's critique of metrics
• Truisms: shall vs. will, what vs. how
• Process evangelism
• Code- and test-centric development
• Hacker's Delusion
• Unified Modeling Language