Note
Notes on reading machines
Half-formed thoughts on why old machine tools are easier to understand than new software, and what that says about both.
A 1960s lathe explains itself. You can stand in front of it and reason your way to what every lever does, because the lever is mechanically connected to the thing it changes. The causal chain is visible.
Modern software has no levers. It has settings, which are claims about behaviour rather than causes of it, and when the claim and the behaviour disagree, there is nowhere to look.
I do not think this is nostalgia. I think legibility is a real property, we used to get it for free from physics, and now we have to design it in deliberately, and mostly we do not bother.
Unfinished thought. Filing it here so I stop losing it.