I've been focusing a lot recently on refactoring. Simultaneously one of the most valuable and rarest developer skills.

In this post, I talk about the different levels of refactoring:

* Primitive (the basic "moves")
* Tactical
* Strategic

codemanship.wordpress.com/2025/12/18/r...

Modularity is the key to fast feedback cycles, short delivery lead times, reliable products and a sustainable pace of innovation.

Apart from that, it's not very important, really.

codemanship.wordpress.com/2025/11/20/m...

When I see dev working backwards from technical strategy to business goals, it feels a bit like writing code and than asking "Now, what test would this pass?"

Write code to make being human easier. Rip code out if it makes being human harder. Write code to make caring for each other easier. Rip it out if it makes caring for each other harder.

Empathy driven development has literally never failed me. It’s done me better than everything else combined

Once you've embraced software development as an iterative, goal-seeking process, it makes most sense to optimise your development process for LEARNING.

And that means optimising for FEEDBACK, and for ease of CHANGE.

It's the feedback loop that does the heavy lifting.

That's to say, teams that struggle with refactoring skills are also teams that most likely struggle with design skills, because tapping into a rich pool of domain and design insights generated as you refactor in small, safe steps is out of their reach.

2/2

Programming is just an exercise in iterating closer and closer to a correct and precise description of the requirements. Good programmers are the ones who are good at guessing what the customer actually wants, so they can skip some of the endless game of telephone!

I wrote this a while ago, and it's still true in many organizations:

"Nobody will validate that the process is actually helpful, because measuring the ongoing investment seems an act of sedition. The process IS. It is unquestioned. It IS because it MUST Be."

medium.com/pragmatic-pr...

A fascinating study of OSS developers found that, while they *reported* significant productivity gains (+20%) using "A.I." coding assistants, the reverse was actually true (-19%).

CASE, 4GLs, Offshoring, Low/No-Code etc. How many times are we gonna fall for this *feeling*?

metr.org/Early_2025_A...

“Performing” teams aren’t teams without conflict. They are teams that have so much practice with conflict they make navigating it look easy.