The "Lisp" programming language covers a large territory, and often one needs to embed some sort of interpreter or language is a complex system. (Even our text

The "Lisp" programming language covers a large territory, and often one needs to embed some sort of interpreter or language is a complex system. (Even our text editors contain sophisticated interpreters because we desire to customize them to such an extent.) Furthermore, a Lisp model is one of the simplest ways to build a powerful interpreter. Consider the problem of being locked in a room that has a computer and all the manuals you need to program it...in machine language. How would you build a system of any significance? Could you escape the evil genie that won't release you until you write a program that solves Sudoku puzzles? I know that I would escape the genie. I'd build a simple macro system and key it into the machine in binary (I've had to patch a machine's boot loader more than once from a panel of switches). Then I'd build a simple assembler from the macro system. After that, I'd build a simple lisp-like system because that's about the simplest language one can implement that does anything interesting, including writing a sudoku solver. So yes, I'd be reinventing Lisp. However, I don't want to spend my life programming in some shitty pathetic system I wrote to escape from an evil genie. Why does it matter that Lisp is so easy to implement that undergrads often build Lisp systems as programming assignments? That doesn't make it some language of the future invented in the past; it just a simple to implement language invented in the past. A thread I've seen here before is "well Algol 68 was a great language and we are not using it because ...<insert some conspiratorial reason>..." Well, Algol 68 is a great language from the past, but it's great not because it's a great language to program in--it's great because it manifested some of the earliest thinking about what features ought to drive the design of programming languages. Today, people aren't locked up and given terrible programming assignments with no real programming language to use (well...maybe I should say this doesn't happen all that often). So why not just embed Lua. Places that I would have used Lisp 40 years ago are now the places I'd use Python or Ruby. But what about the meta-object protocol? What about programs that need to write other programs? What about DSL's? Well, that's another subject. But the short answer is so what? These step up the power of the language one programs in, but not essentially the limits to what a program in your language is capable of doing. Yes, writing complex data structures in old FORTRAN is a big pain, but we aren't in that arena anymore. Writing complex data structures in C++ or Java or Go or Rust isn't a big pain anymore.