Abstract
Syntax Trees (ASTs) and Class Dependency Networks (CDN) constructed from source code. For ASTs, the approach extracts symbolic tokens and encodes them into vectors. These vectors are inputs to a CNN that cap tures semantic information. For CDN, the approach uses a Graph Con volutional Network (GCN) to automatically learn structural information about the network. The approach then combines the semantic and structural information, along with traditional hand-crafted features, to train a classifier that can provide accurate defect prediction. Experi ments using 21 open-source projects show the proposed method out performs state-of-the-art defect prediction models for both withinproject prediction and cross-project prediction. The third paper, Guidelines for the development of a critical software under emergency by Andrea Bombarda, Silvia Bonfanti, Cristiano Gal biati, Angelo Gargantini, Patrizio Pelliccione, Elvinia Riccobene, and https://doi.org/10.1016/j.infsof.2023.107227 Available online 5 April 2023 0950-5849/© 2023 Published by Elsevier B.V. Masayuki Wada, describes a case study on the certification of healthcare devices under emergency situations. The case study includes the in spection and analysis of a comprehensive set of artifacts, including source code, code comments, change requests, test reports, deliverables produced for the software certification (safety concepts, requirements specifications, architecture and design, testing activities, etc.), notes, whiteboard sketches, and emails. The paper includes a number of interesting findings, especially in terms of lessons learned and guidelines for experts, about certifying software for healthcare devices produced with strict and pressing time constraints and with the difficulty of establishing a heterogeneous development team made of volunteers. The fourth paper, Understanding and predicting incident mitigation time by Weijing Wang, Junjie Chen, Lin Yang, Hongyu Zhang, and Zan Wang, proposes TTMPred, a deep-learning-based approach for Time To Miti gation (TTM) prediction in the continuous triage scenario. The paper includes both an empirical study, designed to reveal information about incidents, a