Narrow Results

Software developments involving multiple organizations such as Open Source Software (OSS)-based projects tend to have numerous defects when one organization develops and another organization edits the program source code files. Developments with complex file creation, modification history (origin), and software architecture (functional layer) are increasing in OSS-based development. As an example, we focus on an Android smart phone and a VirtualBox development project, and propose new visualization techniques for product metrics based on file origin and functional layers. One is the Metrics Area Figure, which can express duplication of edits by multiple organizations intuitively using overlapping figures. The other is Origin City, which was inspired by Code City. It can represent the scale and other measurements, while simultaneously stacking functional layers as 3D buildings. The contributions of our paper are to propose new techniques, implement them as web applications, and share the results of our questionnaire. Our proposed techniques are useful not only to visualize the measured metrics, but also to improve the product quality.

This paper presents a novel quality model, which provides a quantitative assessment of the attributes evaluated at each stage of development of scientific applications. This model is defined by selecting a set of attributes and metrics that affect the quality of applications. It is based on the established quality standards. The practical application and verification of the quality model is confirmed by two case studies. The first is an application for solving one-dimensional and two-dimensional Schrödinger equations, using the discrete variables representation method. The second is an application for calculating an ECG-derived heart rate and respiratory rate. The first application follows a development model for scientific applications, which includes some software engineering practices. The second application does not use a specific development model, rather, it is developed ad hoc. The quality of the applications is evaluated through comparative analyses using the proposed model. Based on software quality metrics, the results of this study indicate that the application for solving one-dimensional and two-dimensional Schrödinger equations produces more desirable results.

The cylindrically symmetric static manifolds are classified for their Petrov types and metrics. This classification besides verifying the earlier result that such manifolds cannot be of petrov type II, III and N, gives a complete list of all static cylindrically symmetric metrics of Petrov type O. In the case of Petrov type D metrics, the results appear as three independent classes metrics.

Given the rapidly expanding capabilities of generative AI models for radiology, there is a need for robust metrics that can accurately measure the quality of AI-generated radiology reports across diverse hospitals. We develop ReXamine-Global, a LLM-powered, multi-site framework that tests metrics across different writing styles and patient populations, exposing gaps in their generalization. First, our method tests whether a metric is undesirably sensitive to reporting style, providing different scores depending on whether AI-generated reports are stylistically similar to ground-truth reports or not. Second, our method measures whether a metric reliably agrees with experts, or whether metric and expert scores of AI-generated report quality diverge for some sites. Using 240 reports from 6 hospitals around the world, we apply ReXamine-Global to 7 established report evaluation metrics and uncover serious gaps in their generalizability. Developers can apply ReXamine-Global when designing new report evaluation metrics, ensuring their robustness across sites. Additionally, our analysis of existing metrics can guide users of those metrics towards evaluation procedures that work reliably at their sites of interest.