1 January 2008 in Computer Standards & Interfaces
MDMa Jesús Fernández Díaz
13 April 2020 in IEEE Access
Although source code programs are commonly written as textual information, they enclose syntactic and semantic information that is usually represented as graphs. This information is used for many different purposes, such as static program analysis, advanced code search, coding guideline checking, software metrics computation, and extraction of semantic and syntactic information to create predictive models. Most of the existing systems that provide these kinds of services are designed ad hoc for the particular purpose they are aimed at. For this reason, we created ProgQuery, a platform to allow users to write their own Java program analyses in a declarative fashion, using graph representations. We modify the Java compiler to compute seven syntactic and semantic representations, and store them in a Neo4j graph database. Such representations are overlaid, meaning that syntactic and semantic nodes of the different graphs are interconnected to allow combining different kinds of information in the queries/analyses. We evaluate ProgQuery and compare it to the related systems. Our platform outperforms the other systems in analysis time, and scales better to program sizes and analysis complexity. Moreover, the queries coded show that ProgQuery is more expressive than the other approaches. The additional information stored by ProgQuery increases the database size and associated insertion time, but these increases are significantly lower than the query/analysis performance gains obtained.
15 August 2009 in Science of Computer Programming
13 August 2007 in IEEE Transactions on Education
A course in compiler construction seeks to develop an understanding of well-defined fundamental theory and typically involves the production of a language processor. In a graduate degree in software engineering, the development of a compiler contributes significantly to the developer's comprehension of the practical application of theoretical concepts. Different formal notations are commonly used to define type systems, and some of them are used to teach the semantic analysis phase of language processing. In the traditional approach, attribute grammars are probably the most widely used ones. This paper shows how object-oriented design patterns represented in unified modeling language (UML) can be used to both teach type systems and develop the semantic analysis phase of a compiler. The main benefit of this approach is two-fold: better comprehension of theoretical concepts because of the use of notations known by the students (UML diagrams), and improvement of software engineering skills for the development of a complete language processor.
1 May 2004 in Journal of Systems and Software
3 April 2013 in IEEE Software
The Java 7 platform includes the invokedynamic opcode in its virtual machine, a feature that lets programmers define-and dynamically change-the linkage of method call sites, thereby maintaining platform optimizations. A comprehensive evaluation of a new library's performance includes a description of how to optimize real Java applications.