The Design and Evaluation of an Interoperable Translation System for Object-Oriented Software Reuse.

Abstract

In this paper, we address the problem of defining a source-to-source translation system for reusable software components. We design an interoperable language for writing software components, and present a system to translate components written in the interoperable language to a set of compatible target languages. We analyze the common features in a set of popular programming languages to inform the design of our interoperable language. We evaluate the utility of our system by using our source-to-source translator to convert two well-known open source Java libraries to C++ and Python, and assess the accuracy and performance of the resulting translation.

Improving mobile app development using transpilers with maintainable outputs

“…Researchers have been studying transpilers at least since the 1980's, with the works of Wallis [38] and Albrecht et al [2] using Ada. More recently, some studies have also focused on different uses for transpilation, which include performing compiler optimizations [39], translating between more than two languages with similar paradigms, [3,35] and creating new languages that ease correctness proofs for programs [32].…”

“…This is the case with J2ObjC. Schaub and Malloy notably developed a semantics-preserving transpiler between Java, C++ and Python that aimed to produce readable code [35]. Their study highlights the importance of choosing input and output languages with compatible paradigms and semantics, noting that when "source language features are used [ Schaub and Malloy also mention that "Ideally, a [...] translation should map standard library calls for the source language to the native standard library calls for the target language, rather than introducing an interoperable library layer" [35, p. 6].…”