I disagree with your assessment, and I guess with Sussman's appraisal as well.
I believe that you are mistaken in your belief that Lisp isn't the most advanced language. The submitted thread contains many comments that state that Lisp is basically a pure implementation of the untyped lambda calculus. I don't think this is correct -- Lisp is itself a unique model of computing, and it generally unifies Turing's and Church's computational model in a relatively unbiased manner IMO. For that reason, I regard Lisp as the most flexible languages of all "multi-paradigmed" languages. The Lisp machine's software was basically exclusively implemented in Lisp; the fact that developers could comfortably accomplish that testifies to Lisp's flexibility.
In Lisp, entire linguistic constructs can be dynamically created to iteratively solve a problem. That's basically a bunch of buzzwords, but
I think the statement holds significance. I've read comments elsewhere regarding Lisp as the "Maxwell's equations" of PLT. I agree with that. If you're not familiar with the meta-circular evaluator, I encourage you to read up on it, as it justifies this statement.
That being said, Haskell's type system is really fantastic. I'd love an implementation of a Lisp with a Hindley Milner-esque type system.
Basically, I think Lisp is the most advanced language because it captures what a language is perfectly. This makes it free to involve and embrace new ideas, and has contributed to its continual success and long life.
> "Lisp is itself a unique model of computing, and it generally unifies Turing's and Church's computational model in a relatively unbiased manner IMO. For that reason, I regard Lisp as the most flexible languages of all "multi-paradigmed" languages."
I agree that Lisp is a flexible language, what I disagree with is that this automatically makes it the 'ultimate' language. By ultimate language I mean the best language we can ever create.
The reason I can say that is I see Lisp's flexibility as a double-edged sword. On one hand you can adapt Lisp to match a task, effectively creating a DSL for the particular application you're working on. On the other hand, the flexibility means that you cannot automatically assume as much about an application structure compared to some other languages.
This point about application structure is probably easiest to explain through the lens of type systems. Some of the benefits that type systems can give to other languages do not necessarily apply to Lisp. Do any Lisps have type systems? Sure they do, but they're always optional. It is this that causes an issue. Consider if you were writing some application that used a Lisp type system, and now you want to pull in an external library that performs a set of functions that it would be non-trivial to implement. There are no guarantees that this external library has any types you can query, regardless of how you designed your own code.
Whilst I appreciate that some Lisp coders are comfortable with 'rolling their own' when it comes to reusable libraries, I'm sure you can think of libraries that would be a lot of work to put together. For example, a library on natural-language processing.
Lisp has its advantages, but I think we still have room to improve when it comes to programming languages. Even if Lisp can express everything that can be calculated, there are other factors to consider when looking at the merits of programming languages.
Edit: I submitted this reply an hour or so after opening the tab, during which interval several others mentioned Shen (which is good, as it deserves a bit of exposure).
Shen (in particular its type system) is an executable Sequent Calculus system [0] written with a small portable lisp. This derives from a propositional logic system, not H-M over typed lambda calculus.
Thank you for clarifying. Can you comment on the relative expressive power of Shen's type system versus e.g. Haskell? I understand that the former can accept as well-typed terms the other can't; in what way is this a consequence of modeling Shen's type system using the Sequent Calculus versus Hindley-Milner deduction?
The page describing Shen uses the terms "proprietary" and "closed source" as if they're good things when describing the "Shen Professional" dialect. (Compiler? Environment? It's a bit opaque.)
I'm certainly no Free Software zealot, but that's been an almost unfailing recipe for getting a language and its ideas ignored for 20+ years now.
And the creators appear to be reserving some features for the "commercial" version, and also attempting to heavily promote that version in a way which—I think, based on a few decades' experience—will actively turn away the people who might otherwise be interested in using it.
I believe that you are mistaken in your belief that Lisp isn't the most advanced language. The submitted thread contains many comments that state that Lisp is basically a pure implementation of the untyped lambda calculus. I don't think this is correct -- Lisp is itself a unique model of computing, and it generally unifies Turing's and Church's computational model in a relatively unbiased manner IMO. For that reason, I regard Lisp as the most flexible languages of all "multi-paradigmed" languages. The Lisp machine's software was basically exclusively implemented in Lisp; the fact that developers could comfortably accomplish that testifies to Lisp's flexibility.
In Lisp, entire linguistic constructs can be dynamically created to iteratively solve a problem. That's basically a bunch of buzzwords, but I think the statement holds significance. I've read comments elsewhere regarding Lisp as the "Maxwell's equations" of PLT. I agree with that. If you're not familiar with the meta-circular evaluator, I encourage you to read up on it, as it justifies this statement.
That being said, Haskell's type system is really fantastic. I'd love an implementation of a Lisp with a Hindley Milner-esque type system.
Basically, I think Lisp is the most advanced language because it captures what a language is perfectly. This makes it free to involve and embrace new ideas, and has contributed to its continual success and long life.