CodeScene

Another reason to tackle technical debt. Pay it down and reduce developer churn. Bad code is a top reason why developers quit... leading to knowledge loss and costly onboarding to replace them - and possibly more debt!

Are we focusing AI on the right problem? If AI helps us generate code faster, what does that mean for quality and maintainability? This Friday, we have something special for you. 🚀 Our Founder and CTO, Adam Tornhill, recently presented an inspiring talk at Codecamp, and for those who couldn’t attend, we’ve put together some key slides. Enjoy! 🎉 #aicoding #codequality #technicaldebt #ai #softwareengineering

What's the true cost of change in software? Dependencies can slow down progress and kill the joy of programming faster than you can say "Visual Basic." But managing dependencies is hard because they are neither inherently good nor bad -- context is everything. If we focus solely on static dependencies -- the ones we see directly in the code -- we will miss out on key information: 1. The most important dependencies are often logical rather than physical. For example, two pieces of code might be deeply intertwined even if they communicate through an intermediary, like a message bus. 2. Our dependency graph won't be actionable due to its overwhelming complexity. In any non-trivial system, there are countless dependencies, including stable elements like standard library functions. So, what if we could evaluate our actual modifications to the code against the desired change patterns? This concept is known as Change Coupling. It's based on a simple fact: code changes for a reason; we don't just change code at random. (If we do, well, we might have bigger problems). As such, Change Coupling isn't measured from code but from the evolution of your code. By tapping into version-control history, you identify clusters of files which are frequently modified as part of the same commit sets. Files that co-evolve are change coupled. You then evaluate those actual change patterns against your architectural principles to catch unexpected and problematic dependencies. At its core, software architecture is about making change easy, whether for bug fixes, new features, or evolving non-functional requirements. The real question is: Does your architecture work with or against you?

Last week, our principal researcher Markus Borg gave two talks at the 40th Conference on Software Maintenance and Evolution in Flagstaff, Arizona (ICSME 2024). What did he share? 🚀 In the first talk, we shared a competitive benchmarking study comparing CodeScene’s Code Health with SonarQube’s maintainability rating. The ground truth was the manually annotated Maintainability Dataset established by Markus Schnappinger and colleagues at the Technical University of Munich. The results were striking: CodeScene significantly outperforms SonarQube in identifying unmaintainable files, with an AUC score of 0.95 vs. 0.60. 🚀 SonarQube’s performance suffers from a high rate of false positives, aligning with previous research findings. You can read the preprint here: https://lnkd.in/dB2pYaMZ. 🚀 The second talk focused on a registered report – the design of a study we’re conducting in collaboration with Equal Experts and Lund University. We’ve structured a randomized control trial to investigate the impact of AI Assistants on software maintainability. Developers, if you're interested in participating, we’d love to hear from you. Reach out to learn more about the next steps! #codequality #aicoding #technicaldebt

Are you attending CodeCamp Festival October 24? Don’t miss the chance to meet our Founder and CTO, Adam Tornhill - one of this year’s speakers! 🚀 Adam will be presenting "Exploring the Fact-based Realities of AI-assisted Coding" https://lnkd.in/eAgXe_cg

Michelin TechDay on Software craftsmanship continues with a lecture from Adam Tornhill on software quality and technical debt :)

In the new issue of IEEE Software, the Requirements column celebrates the magazine’s 40th anniversary 🎉 A milestone! What better way to honor this moment than by reflecting on 40 years of progress in requirements engineering? I’m happy to share a discussion between two esteemed Swedish professors, Kristian Sandahl and Bjorn Regnell. Both offer their personal reflections on the evolution of the field over the past four decades. A fun read for anyone interested in requirements engineering! Check out their thoughts here: 👉 Preprint: https://lnkd.in/dzq57jea 📰 Full paper: https://lnkd.in/dEhepT_Q #requirementsengineering #IEEEsoftware #reqeng

#theEngineeringClub just released a new episode. 🎉Don't miss this amazing podcast with Adam Tornhill about AI's impact on code quality. 👇 #codequality #aicoding #refactoring #ai

Ready for Friday? We've got your knowledge boost ready! This time, we’re diving into AI coding and how to fix code smells in your editor. So when is the best time to address code health issues? As early as possible —while you're still writing the code. 🚀 Join us as we demonstrate how to detect and eliminate code smells directly in your editor. With the Code Health Monitor, you’ll stay informed without workflow interruptions. And if needed, CodeScene ACE’s AI-powered refactoring can automatically fix issues seamlessly. Happy Friday! 🚀

We're inviting software professionals to participate in a short survey, part of a collaborative master’s thesis between CodeScene and Lund University. 🚀 This research, conducted by Martin Larsson and Philip Breid and supervised by Dr. Markus Borg and Peter Anderberg, aims to deepen our understanding of industry practices around software maintainability. 📝 Join the Study Your participation would be greatly appreciated! Click the link below to take the survey and contribute to advancing software maintainability practices:  🔗 https://lnkd.in/dUwU86yc