Micro-Frontends for Mobile Applications and more

Links

• Micro-Frontends applied to mobiles? We’ll see!

  Micro-Frontends is an architecture style primarily suited for web client applications, as it allows the integration of different types of frontends regardless of their underlying technology. However, applying this architecture to mobile clients is not yet possible due to various reasons. Cross-platform solutions have limitations in combining UI components designed with different technologies. Kotlin Multiplatform Mobile (KMM) offers a promising approach by sharing native business logic across multiple clients. Progressive Web Apps (PWA) are web applications that can benefit from micro-frontends, but they are not considered mobile applications. Native technologies provide alternatives for organizing and sharing UI components within their respective scopes. Declarative UIs in iOS and Android enable view composition but are not equivalent to micro-frontends. While a fully agnostic micro-frontend architecture is currently unattainable for mobiles, a modular design with declarative views using a single technology is a more feasible approach.

• Copilot Leaks: Code I Should Not Have Seen

  GitHub Copilot is an impressive tool for generating code, but it should not be blindly trusted. The author shares their experience with Copilot, highlighting instances where it produced code with obvious issues and exposed potentially private information. The tool often suggests publicly available APIs, but occasionally invents fictional or private ones. Leaked API endpoints, keys, and credentials pose a risk, especially when accidentally uploaded to GitHub. Financial APIs are less frequently provided, and proprietary specifications and formats are not accurately implemented.

• 7 Best-In-Class AI Tools Every CTO Needs to Know

  AI is revolutionizing software engineering, and there are several "best-in-class" tools available for AI adoption. GitHub Copilot X is a highly anticipated AI tool that integrates into engineers' workflows, offering features like code explanation, error fixing, unit test generation, and more. Sourcegraph Cody helps developers read, write, and understand code faster, while CollabGPT serves as an AI companion for software projects, providing insightful overviews and strategic suggestions. Readable automates the process of generating code comments, Adrenaline AI offers comprehensive code comprehension, Grit.io automates technical debt management, and Codeball AI enhances the code review process. Embracing AI in software engineering can lead to significant improvements in productivity and efficiency.

• Decoupling a core service from your monolith the right way

  This blog post details the process of decoupling the billing logic from WeTransfer's monolithic application to improve feature shipping and reduce technical debt. The author describes the four phases of the project, starting with decoupling the billing logic within the monolith, then onboarding the new billing service called Billy. They also discuss the use of feature flags and monitoring, followed by cleanup and the eventual database split. The post highlights the benefits achieved, such as code removal, increased productivity, improved availability, and the importance of careful planning and teamwork in such a complex undertaking.

Github repositories

• ChuloAI/oasis

  Oasis is a code generation tool that uses open source LLM models and a guidance library. It is available as a VSCode plugin, providing code generation commands through right-click or the command palette. The latest update (v0.2.0) includes a separate repository for the guidance server code. Setting up Oasis requires cloning the repository, downloading the desired model, and running it with Docker. There are instructions for running on CPU or GPU. The tool allows parsing input code, applying guidance prompts, and generating responses. The plugin has some known issues related to newlines and indentation. More details can be found in the provided Medium article.