Static Analysis of Dynamic Languages
Year: 2015
Type: dissertation
Abstract: Dynamic programming languages are highly popular and widely used. Java- Script is often called the lingua franca of the web and it is the de facto standard for client-side web programming. On the server-side the PHP, Python and Ruby languages are prevalent. What these languages have in common is an expressive power which is not easily captured by any static type system. These, and similar dynamic languages, are often praised for their ease-of-use and flexibility. Unfortunately, this dynamism comes at a great cost: The lack of a type system implies that most errors are not discovered until run-time. Thus, in the worst-case, these bugs are not uncovered before they are encountered by real users of the system. A further cost is limited tool support: For instance, integrated development environments with code completion, code navigation and automatic refactorings are widely available for languages with static type systems, such as Java and C# , but the same features are rarely available for dynamic languages such as JavaScript. The aim of this thesis is to investigate techniques for improving the tool- support for dynamic programming languages without imposing any artificial restrictions on the behaviour of these languages. A common theme is the reliance on static program analysis to over-approximate the behaviour of programs written in these languages. Specifically, the use of whole-program dataflow analysis. The research challenge of this line of work is the adaption of existing- and invention of new dataflow analysis techniques to tackle the nature of dynamic programming languages.
Authors: Magnus Madsen, Phd Dissertation, Magnus Madsen
Institution: Aarhus University
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4412328370
