I shipped the same stack for 30 years on purpose

Counter-consensus: chasing new stacks is not engineering, it is anxiety.

The same Smalltalk-derived line — now the Glamorous Toolkit — has held for thirty years, and it held because it is a living system rather than a frozen vendor product. Every layer of it could be lifted, looked at, and changed by the people using it.

The cost of the new stack is not only the migration. It is the years spent learning a system that will itself be replaced, and the accumulated judgement that goes with it, thrown away. Most of the value in a stack is not in the stack. It is in knowing its sharp edges and where to stand.