summaery2018: Projects

Generating Code Suggestions

Project information

submitted by

André Karge

Co-Authors

Mentors

Prof. Dr.-Ing. Norbert Siegmund

Faculty / Section:
Media

Degree programme:
Computer Science and Media (english) (Master of Science (M.Sc.))

Type of project presentation

Final project

Semester

Winter semester 2017/18

Exhibition Location / Event Location
  • Bauhausstraße 11
    (R015)

Project description online

As a result of his master thesis, André Karge developed a system to generate source code which can be used to help programmers in their software development process.

Most programmers use integrated development environments (IDEs) to develop software.
These environments provide a variety of tools to support the development process, which includes automatic source-code suggestions.

Furthermore, developers frequently include external libraries in their software.
When working with libraries, it could be that many parts of the source code are similar to already existing source code snippets.

The aim of this approach is to use this fact in order to improve the automatic code suggestion system with enhanced suggestions.

To this end, a neural network was developed which uses a variety of source code files to learn such similarities to generate suggestions based on the already written source code.

Exhibition Location / Event Location