To support software development, integrated development environments (IDEs) provide many services, including programmer-friendly feedback while a program is being edited. Typically, these IDEs support not only various programming languages, but also an evolving composition of languages. Thus, unlike traditional parsers, the parsers in IDEs are designed for multiple languages and must be responsive while handling error recovery, syntax highlighting, and content completion.

The approach proposed in this paper addresses these issues effectively. It integrates automatically maintained recovery rules; heuristics; grammar formalisms, such as island and permissive grammars; and parsing techniques, such as region recovery and bridge parsing. The implementation of such a system is tested in several experiments, and the results demonstrate the quality, performance, and scalability of the system. While the authors intended their studies to be comprehensive, it should be noted that test cases are generated in an ad hoc manner. A comparison with an existing IDE (Eclipse Java development tools (JDT)) shows the practicality of the proposed approach. Computer-aided software engineering (CASE) developers and compiler writers would greatly benefit from this proposal.

Even though the contributions to generalized parsing are substantial, the presentation is verbose, lengthy, and without focus. Digressions and repetitive statements distract from the main topic and make reading the paper burdensome. Moreover, unlike expositions that are found in the traditional compiler theory textbooks, the exposition here suffers from a lack of formality. Further research is needed to provide such formality.