Why Lisp is a Big Hack and Haskell is Doomed to Succeed (2011)

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by iopq on Sunday March 30, 2014 @06:04PM (#7498158)

The singular "perfect" language would have to have pluggable syntax - so that anyone can have their perfect syntax as long as it generates the same parse tree.

At the same time, its type system will be pluggable too, so that people can keep improving the type system (dependent types, dynamic types, etc.) without changing the language.

Oh yeah, and of course its code generation/execution model will be pluggable as well. Do you want to interpret it? Sure! Do you want to compile it so it performs better on your target computers? No problem! Compile it to JavaScript? Why not?

In that way everyone can be programming in the same language that's flexible enough for everyone's use, but at the same time can contain DSLs. Pluggable syntax/custom type checking means you can embed SQL code and make sure it's valid before running it. It also means SQL could be a custom SQLStatement data type with its own semantics.

In short, only a language that allows every type of programming can be the "one true language"

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by taeric on Sunday March 30, 2014 @08:41PM (#7498751)

Doesn't this just describe lisp, though? Most of your questions have a dialect or library set that is specifically made to address them.

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by iopq on Sunday March 30, 2014 @10:16PM (#7499149)

No, lisp doesn't allow you to specify your own syntax. You can't just start writing stuff without parens and define what that parses to later.

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by taeric on Sunday March 30, 2014 @11:39PM (#7499472)

You'd be somewhat surprised on that point. Especially if you just take it such that everything is just trying to build up an s-expression. While the homoiconicity of the language is incredibly cool for macros and whatnot, I don't think you strictly need it. Especially not at the top level. (That is, if you made a language that "compiled" down to s-expressions, what is missing?)

At the extreme end, take a look at Dylan.

Though, I was really referring to your other points. It seemed every one of your "questions" is directly addressed.

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by mcv on Monday March 31, 2014 @10:42AM (#7501720)

But everybody defining their own syntax would not make a language "the perfect language". Quite the opposite, because nobody would be able to understand each other's code anymore. It would be the ultimate fragmented language.

And because of that, I don't think a perfect language is even theoretically possible. Everybody has their own syntactical preferences, and allowing all of them means fragmentation.

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by xenophonf on Monday March 31, 2014 @04:11AM (#7500266)

Yes, it does. Look for "reader macros" in the Common Lisp spec, do a search for "lisp infix notation", etc.

[ Parent | on HN ]

Re:Why Lisp is a Big Hack and Haskell is Doomed to Succeed (... by calibraxis on Monday March 31, 2014 @03:54AM (#7500224)

How do you evaluate Common Lisp's reader macros?

[ Parent | on HN ]