Software Maintenance and Evolution

Teacher(s)

Mens Kim;

Language

English

Main themes

Whereas many software engineering courses focus on building new systems from scratch, in industrial practice software developers are often confronted with already existing software systems that need to be maintained, reused or evolved. This requires specific skills to understand the design and implementation of an existing system and which parts need to be modified, to build software systems that are easier to maintain, and to design systems with reuse and evolution in mind from the very start.

This course will thus study a variety of techniques, tools and methodologies to help building software systems that are easier to understand, maintain, reuse and evolve, such as:

Preliminaries and definitions :
               - need for and problems of software maintenance and evolution
               - definitions, differences between and types of software maintenance and evolution
               - technical debt
               - laws of software evolution

Domain modelling :
               - software product lines
               - domain analysis
               - feature modelling
               - commonalities and variabilities
               - feature diagrams

Software reuse :
               - definition of and needs for software reuse
               - reuse techniques and design for reuse
               - object-oriented techniques for reuse and maintainability
               - object-oriented application frameworks

Software maintenance and evolution :
               - Best programming practices and coding standards
               - Code refactoring and reengineering
               - Bad code smells
               - Software and design patterns 
               - Design principles and heuristics

An industrial case study

Learning outcomes

At the end of this learning unit, the student is able to :
1	Given the learning outcomes of the "Master in Computer Science and Engineering" program, this course contributes to the development, acquisition and evaluation of the following learning outcomes: INFO1.1 , INFO1.3 INFO2.5 INFO5.5 Given the learning outcomes of the "Master [120] in Computer Science" program, this course contributes to the development, acquisition and evaluation of the following learning outcomes: SINF1.M3 SINF2.5 SINF5.5 Students completing successfully this course will be able to Understand the difficulties of developing code in a change context as opposed to 'green field' development Assess the impact of a change request to an existing product of medium size. Describe techniques, coding idioms and other mechanisms for implementing designs that are more maintainable. Understand how design patterns can improve the design of a software system. Refactor an existing software implementation to improve some aspect of its design. Identify the principal issues associated with software evolution and explain their impact on the software lifecycle.' Discuss the advantages and disadvantages of different types of software reuse.

Content

The course will cover a variety of techniques, tools and methodologies to help building software systems that are easier to understand, maintain, reuse and evolve.
Preliminaries:

• Definitions and difference between software maintenance, software evolution and software reuse
• Different types of software maintenance and evolution
• Causes for software maintenance and change
• Technical debt
• Laws of software evolution

Domain modelling:

• Domain modelling and domain analysis
• Software product lines
• Feature-oriented domain analysis
• Feature modelling, commonalities and variabilities
• Feature relationships, dependencies and cross-tree constraints
• Semantics of feature models and feature model anomalies

Software reuse:

• Definitions of reusability, software reuse and reusable components
• How object-oriented programming promotes modularity, maintainability and reuse
• Encapsulation, information hiding, polymorphism and code sharing
• Key object-oriented concepts: object, classes, methods, messages, inheritance
• Polymorphism and dynamic binding
• Method overriding, self and super calls
• Abstract classes and methods
• Different kinds of inheritance: single, multiple, interfaces, mixins

Bad code smells:

• Bad smells
• Bad smells vs. refactorings
• Bad smell categories and examples
• Coupling and cohesion

Code refactoring:

• Refactoring (definitions, motivations, when should you refactor)
• Refactoring categories and examples
• Refactoring vs. code quality
• Merge conflicts due to refactoring

Software patterns:

• Christopher Alexander’s building architectural patterns
• (Software) design patterns (definitions, motivations, structure)
• Abstract Factory design pattern
• Factory Method design pattern
• Strategy and Decorator design patterns
• Antipattern (definition, purpose, example: The Blob)
• The 7 deadly sins

Design heuristics:

• Design heuristics (definition, purpose, examples)
• Design heuristics related to inheritance and polymorphism
• Design heuristics related to cohesion
• Design heuristics related to coupling

Application frameworks:

• Object-oriented application frameworks (definition, purpose, examples)
• How frameworks can achieve software reuse
• Inversion of Control (the “Hollywood” principle)
• Software frameworks vs. libraries
• Hotspots and hook methods
• Commonality and variability
• White vs. grey vs. black box frameworks
• Template method design pattern
• Design patterns vs. frameworks
• Refactoring to a framework
• Using template methods to evolve an application into a framework
• Refactoring to specialise or generalise class hierarchies

Industrial case study (invited speech by a selected company)
Additional research-related sessions (if time remains) on selected topics such as:

• Context-Oriented Software
• Reflection and metaprogramming
• Aspect-oriented programming

Teaching methods

Evaluation methods

The course evaluation will be a weighted average of:

Other information

Even though good quality software may be easier to maintain and evolve, software quality assurance techniques will not be addressed explicitly in this course as they are the topic of a separate course on Software Quality Assurance [LINGI2251]
Expected background:

• Having a good knowledge of and experience with object-oriented programming concepts, algorithms and data structures.
• Having prior or simultaneous experience with the development of a medium- to large-scale software system.

Online resources

Moodle course website
The course slides as well as other relevant and practical information related to the course will be accessible on Moodle. The same platform will also be the means of communication between the teacher(s) and the students.

Bibliography

English

Faculty or entity

INFO