Published on 23 May 2014.
I remember thinking many years ago that text is a clumsy way to work with code. Why do we operate at the level of characters when we are editing code that has a more well defined structure?
Automatic refactoring tools raise the level a bit. With those, we can think in terms of inline variable and have that be a single operation in our editor. There is no gap between what we want to do and how we do it. If we are just working with text, inline variable involves many steps, some of which are delete occurrence of variable, copy and paste the assigned value to those places, delete the declaration.
But refactorings limit us to the catalogue implemented by the tool. The other operations we want to do, we have to do by editing at the character level.
Yesterday I saw an Emacs plugin that let you modify Lisp expressions at a higher level. It was a plugin for modifying the AST directly. (Or at least it felt like that.) It feels like an AST editor could be more powerful than a text editor. Perhaps it would be easy to define our own editing commands/shortcuts in terms of operations to an AST?
Site proudly generated by Hakyll.