Publication Date

Spring 2017

Degree Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Engineering

Advisor

Mohamed Fayad

Keywords

analysis, design, pattern, reuse, software, stability

Subject Areas

Computer engineering; Computer science; Systems science

Abstract

Software is a necessity in the modern world, and that need is continuously growing. As expensive as the creation of all this new software is, the maintenance costs are even greater. One solution to this problem is software reuse, whereby already written software can be applied to new problems after some modification, thus reducing the overall input of new code. The goal in traditional software reuse is to produce a piece of software with enough flexibility to be used at least twice. Unfortunately, there are many difficulties in achieving software reuse using modern programming techniques. Even software built specifically for reuse is severely restricted in its utility for new applications. It is easy for new programs to require entirely new logic or new objects. Because of this, they become quickly outdated, and any labor spent creating reusable software is nullified. The solution is a method to vastly increase the reusability of software by concentrating on the base knowledge and overall goals of software rather than the details on a case-by-case basis. Finding patterns in the problem and solution spaces allows unification into a smaller solution set. Instead of each problem receiving its own solution from marginally reusable components, multiple problems are resolved with the same architecture and object set. As an added benefit, this solution will not only vastly improve software reuse, but it will make feasible systems that can construct software architecture on demand and provide the first steps to fully automated software development.

Share

COinS