Ooh Ooh My Turn Why Lisp? (2008)

Post by ruste » Tue, Aug 02, 2016, 7:36 PM UTC

Frondo wrote: post #116

snaky wrote: post #60

Earlier quoted context omitted.

> doing anyone any good There's a difference between doing and talking about. Compare an amount of noise from Rust crowd to anything useful being done in Rust. Let's look at only one product implemented in Lisp, by only one vendor, and it's users in only one industry - http://allegrograph.com/healthcare/ . Pfizer, Mayo Clinic, GSK, Novartis - should we stop with that 'doing anyone any good' mantra?

The thing is, I'm trying to ask, if lisp is super-powerful and a secret weapon (the general sense of the claims) why isn't the real world littered with examples of companies who seriously out-competed using Lisp? So far all people can do is say "no, SOMEONE is using Lisp!" Well, great, but that's not the claim from the Lisp community. The claim is that Lisp is powerful, so powerful that it's a secret weapon. Why is t…

Post by groovy2shoes » Tue, Aug 02, 2016, 8:25 PM UTC

emidln wrote: post #140

groovy2shoes wrote: post #138

Earlier quoted context omitted.

The reason typing is not a solved problem is this: there are valid programs which can be expressed in an untyped language that cannot be (directly) expressed in a typed one at the moment. For example: (define (foo p?) (if p? 42 "forty-two")) What is the type of `foo`? It could be `bool -> int` or `bool -> string`, depending on the result of `p?`! In an untyped language, this is a perfectly valid program. In a typed l…

In Shen, we easily typecheck foo to the specific values 42 and "forty-two", not merely either String or Int. \\ enforce type checking (tc +) \\ create a forty-two type (datatype forty-two _______________ 42 : forty-two; ________________________ "forty-two" : forty-two;) \* example (7+) 42 42 : number (8+) 42 : number 42 : forty-two (9+) 43 43 : number (10+) 43 : forty-two type error *\ \\ define our foo that takes a…

Post by emidln » Tue, Aug 02, 2016, 10:41 PM UTC

groovy2shoes wrote: post #143

emidln wrote: post #140

Earlier quoted context omitted.

In Shen, we easily typecheck foo to the specific values 42 and "forty-two", not merely either String or Int. \\ enforce type checking (tc +) \\ create a forty-two type (datatype forty-two _______________ 42 : forty-two; ________________________ "forty-two" : forty-two;) \* example (7+) 42 42 : number (8+) 42 : number 42 : forty-two (9+) 43 43 : number (10+) 43 : forty-two type error *\ \\ define our foo that takes a…

Interesting! Such an extensional definition of a type could be really flexible, but seems like it would be limited to finite types. Is this the case? Or can you specify types with arbitrary judgments?

    (datatype odd-integer
      if (integer? X)
      if (odd? X)
      ________________
      X : odd-integer;)

Post by junke » Wed, Aug 03, 2016, 9:16 AM UTC

groovy2shoes wrote: post #138

effie wrote: post #123

Earlier quoted context omitted.

"Typing is not a solved problem, because if it was, there wouldn't be a static vs dynamic type debate that lingers ad nauseum." Could you clarify? Surely even if typing was completely solved, we would still have debates of that kind, because there are benefits to both approaches and there will always be people preferring one or another.

The reason typing is not a solved problem is this: there are valid programs which can be expressed in an untyped language that cannot be (directly) expressed in a typed one at the moment. For example: (define (foo p?) (if p? 42 "forty-two")) What is the type of `foo`? It could be `bool -> int` or `bool -> string`, depending on the result of `p?`! In an untyped language, this is a perfectly valid program. In a typed l…

    (FUNCTION (T)
      (VALUES
        (OR (INTEGER 42 42)
            (SIMPLE-ARRAY CHARACTER (9)))
       &OPTIONAL))

Post by gkya » Wed, Aug 03, 2016, 11:10 AM UTC

Frondo wrote: post #116

snaky wrote: post #60

Earlier quoted context omitted.

> doing anyone any good There's a difference between doing and talking about. Compare an amount of noise from Rust crowd to anything useful being done in Rust. Let's look at only one product implemented in Lisp, by only one vendor, and it's users in only one industry - http://allegrograph.com/healthcare/ . Pfizer, Mayo Clinic, GSK, Novartis - should we stop with that 'doing anyone any good' mantra?

The thing is, I'm trying to ask, if lisp is super-powerful and a secret weapon (the general sense of the claims) why isn't the real world littered with examples of companies who seriously out-competed using Lisp? So far all people can do is say "no, SOMEONE is using Lisp!" Well, great, but that's not the claim from the Lisp community. The claim is that Lisp is powerful, so powerful that it's a secret weapon. Why is t…