Got my copy of 'Your Code as a Crime Scene' (2nd edition) by Adam Tornhill in the mail today. Tools like CodeScene (and more broadly, the techniques discussed in the book) will become even more important as AI-generated code becomes more prevalent.

One reason AI works for code is that most people are just writing the same programs over and over. The elegant solutions to this problem are more abstract languages and more powerful libraries. But maybe AI will be the worse-is-better solution that wins.

We're very proud to have Markus Borg Adam Tornhill, and enys mones 🇺🇦 present their technical paper 'Increasing, not Diminishing: Investigating the Returns of Highly Maintainable Code' on TechDebtConf 2024 🔥
conf.researchr.org/program/TechDe…

Credit to co-writer Ilyana Pruvost as well 💙

~”Instead of trying to measure developer productivity, via highly gameable metrics, measure the levels of waste in the system”

— Scott Shaw, DevOpsDays Singapore

Test code frequently has technical debt, too.

One pattern that always worries me are logical dependencies between different test suites. Why?
Because you make a tiny change to the application code, and now you have to play shotgun surgery hunting down broken test cases.

Under-engineered code is a more common problem than over-engineering.

Living in Gothenburg and working within software industry? Don't miss our workshop March 26 with our CTO Adam Tornhill

🔥Sign up: hubs.ly/Q02pX-wJ0

Brilliantly precise talk by Adam Tornhill on the business impact of code quality. youtu.be/ty9d7wEVTwc

Twenty years ago, Kevlin Henney's work inspired me to start writing about software myself. I remain a big fan of Kevlin. That’s why I’m extra proud to have his foreword in the new edition of Your Code as a Crime Scene.

Attending the <Programming> conference in Lund this week. I've presented a demo of our LLM-based refactoring already, but I'll be around to let anyone explore it further. Feedback much appreciated!

The question of whether Machines Can Think...is about as relevant as the question of whether Submarines Can Swim.
-- Edsgar Dijkstra

VoxxedDays Zurich opens its doors this Thursday at 8:00 am and kicks off the day at 9:00 am with the welcome and keynote by Adam Thornhill. Make sure you don't miss out 🚀🎁🥰🙏🧑‍💻

voxxeddays.com/zurich/

#VDZ24 #voxxed #voxxed days #itwillbefun

Code is read much, much more frequently than it's written. Being quick at spotting problematic coding constructs is a valuable skill. Such an important training video. 👍

Asking 'how long should a function be' in terms of lines of code is IMHO the wrong question. Focus on cohesive and well-named functions instead, which aligns with the way our brain prefers to view code.

Here's how I tried to put it in Your Code as a Crime Scene (2nd ed):

I'm honoured to announce that I will be giving the keynote at Voxxed Days Zürich next week on How Empirical Data Shatters the Speed vs Quality Myth.

Based on novel research, we'll see that it's indeed possible to move fast *without* breaking things. I'm looking forward to it!

I'm super proud of these results: the benchmarking is a great validation of all the work we've been doing at building a more reliable code quality metric.

It’s easy to be on top when you’re the only one 😂

Did you know that we developers spend 70% of our time on understanding existing code?
This means that if we want to optimize any aspect of software development, we should optimize for ease of understanding. That's the big win.
youtube.com/watch?v=dXY6Fy…