Recent research provides evidence that effective communication in collaborative software development has significant impact on the software development lifecycle. Although related qualitative and quantitative studies point out textual characteristics of well-formed messages, the underlying semantics of the intertwined linguistic structures still remain largely misinterpreted or ignored. Especially, regarding quality of code reviews the importance of thorough feedback, and explicit rationale is often mentioned but rarely linked with related linguistic features. As a first step towards addressing this shortcoming, we propose grounding these studies on theories of linguistics.We particularly focus on linguistic structures of coherent speech and explain how they can be exploited in practice.We reflect on related approaches and examine through a preliminary study on four open source projects, possible links between existing findings and the directions we suggest for detecting textual features of useful code reviews.@en

The software development process produces vast amounts of textual data expressed in natural language. Outcomes from the natural language processing community have been adapted in software engineering research for leveraging this rich textual information; these include methods and readily available tools, often furnished with pre-trained models. State of the art pre-trained models however, capture general, common sense knowledge, with limited value when it comes to handling data specific to a specialized domain. There is currently a lack of domain-specific pre-trained models that would further enhance the processing of natural language artefacts related to software engineering. To this end, we release a word2vec model trained over 15GB of textual data from Stack Overflow posts. We illustrate how the model disambiguates polysemous words by interpreting them within their software engineering context. In addition, we present examples of fine-grained semantics captured by the model, that imply transferability of these results to diverse, targeted information retrieval tasks in software engineering and motivate for further reuse of the model.@en

Context: An excessive number of code smells make a software system hard to evolve and maintain. Machine learning methods, in addition to metric-based and heuristic-based methods, have been recently applied to detect code smells; however, current methods are considered far from mature. Objective: First, explore the feasibility of applying deep learning models to detect smells without extensive feature engineering. Second, investigate the possibility of applying transfer-learning in the context of detecting code smells. Methods: We train smell detection models based on Convolution Neural Networks and Recurrent Neural Networks as their principal hidden layers along with autoencoder models. For the first objective, we perform training and evaluation on C# samples, whereas for the second objective, we train the models from C# code and evaluate the models over Java code samples and vice-versa. Results: We find it feasible to detect smells using deep learning methods though the models’ performance is smell-specific. Our experiments show that transfer-learning is definitely feasible for implementation smells with performance comparable to that of direct-learning. This work opens up a new paradigm to detect code smells by transfer-learning especially for the programming languages where the comprehensive code smell detection tools are not available.@en

The emergence of online open source repositories in the recent years has led to an explosion in the volume of openly available source code, coupled with metadata that relate to a variety of software development activities. As an effect, in line with recent advances in machine learning research, software maintenance activities are switching from symbolic formal methods to data-driven methods. In this context, the rich semantics hidden in source code identifiers provide opportunities for building semantic representations of code which can assist tasks of code search and reuse. To this end, we deliver in the form of pretrained vector space models, distributed code representations for six popular programming languages, namely, Java, Python, PHP, C, C++, and C#. The models are produced using fastText, a state-of-the-art library for learning word representations. Each model is trained on data from a single programming language; the code mined for producing all models amounts to over 13.000 repositories. We indicate dissimilarities between natural language and source code, as well as variations in coding conventions in between the different programming languages we processed. We describe how these heterogeneities guided the data preprocessing decisions we took and the selection of the training parameters in the released models. Finally, we propose potential applications of the models and discuss limitations of the models.@en