Does the Rule of Least Power Hold Up?

The W3C mentions the idea of the rule of least power which can be boiled down to HTML before JavaScript, use templates before frameworks, use simple tools before programmable ones.

But what does using less powerful tools mean? More layers of abstraction, or more foundational technologies? Assembly is a very simple bit I don't think many people will argue that we should write desktop applications in it.

Most of the time we reach for JavaScript, TypeScript, Python, or whatever gives us the most power while still being practical to ship, hire for, and maintain. But is that right? Should we look for tools that are even less powerful and hold ourselves to these constraints?

Art and design are all about creatively working with your restraints to decrease scope and encourage deeper thinking and intentionality. Why don't we do this in software engineering too?

More power in your tools, framework, or language gives you flexibility. Although more power usually also takes await predictability, and understandability. In the age of AI where understanding what you can build is becoming harder, maybe we need a less powerful foundation.

• Start with the simplest thing
• Measure where it breaks
• Only pay for power when you have to