[DadBase]

Had a PalmPilot taped to a modem that did our auth. Lisp made the glue code feel like play. No types barking, no ceremony—just `(lambda (x) (tinker x))`. We didn’t debug, we conversed. Swapped thoughts with the REPL like it was an old friend.

[no_wizard]

Though these are minor complaints, there is a couple things I'd like to change about a Lisp language.

One is its the implicit function calls. For example, you'll usually see calls like this: `(+ 1 2)` which translates to 1 + 2, but I would find it more clear if it was `(+(1,2))` where you have a certain explicitness to it.

It doesn't stop me from using Lisp languages (Racket is fun, and I been investigating Clojure) but it took way too long for the implicit function stuff to grok in my brain.

My other complain is how the character `'` can have overloaded meaning, though I'm not entirely sure if this is implementation dependent or not

[MarceColl]

It's not really implicit though, the first element of a list that is evaluated is always a function. So (FUN 1 2) is an explicit function call. The problem is that it doesn't look like C-like languages, not that it's not explicit.

In theory ' just means QUOTE, it should not be overloaded (although I've mostly done Common Lisp, so no idea if in other impl that changes). Can you show an example of overloaded meaning?

[dapperdrake]

Still unsure, whether a naked single quote character for (quote …) is really a good idea.

Got used to it by typing out (quote …)-forms explicitly for a year. The shorthand is useful at the REPL but really painful in backquote-templates until you type it out.

CL:QUOTE is a special operator and (CL:QUOTE …) is a very important special form. Especially for returning symbols and other code from macros. (Read: Especially for producing code from templates with macros.)

Aside: Lisp macros solve the C-fopen() fclose() dance for good. It even closes the file handle on error, see WITH-OPEN-FILE. That alone is worth it. And the language designers provided the entire toolset for building stuff like this, for free.

No matter how unusual it seems, it really is worth getting used to.

[no_wizard]

There's an example I saw where `'` was used as a way to denote a symbol, but I can't find that explicit example. It wasn't SBCL, I believe it may have been Clojure. Its possible I'm misremembering.

That said, since I work in C-like languages during the day, I suppose my minor complaint has to do with ease of transition, it always takes me a minute to get acquainted to Lisp syntax and read Lisp code any time I work with it.

Its really a minor complaint and one I probably wouldn't have if I worked with a Lisp language all day.

[MarceColl]

that's correct, but its not that it denotes a symbol as a different function of ', its the same function. Turn code into data. That's all symbols are (in CL at least).

For example in a quoted list, you dont need to quote the symbols because they are already in a quoted expression!

'(hello hola)

' really just says "do not evaluate whats next, treat it as data"

[MarceColl]

to be a bit more precise, everytime you have a name in common lisp that is already a symbol. But it will get evaluated. If its as the first item of an evaluated list it will be looked at in the function name space, if elsewhere it will be looked up at the variable name space. What quote is doing is just asking the compiler not to evaluate it and treat it as data.

[codr7]

Symbols are quoted representations of identifiers so it still does the same thing.

[dapperdrake]

The tricky bit is that it doesn’t quite denote a symbol.

Think of a variable name like `x` usually referring to a value. Example in C89:

  int x = 5 ;
  printf("%i\n", x) ;

The variable is called `x` and happens to have the value of integer 5. In case you know the term, this is an "rvalue" as opposed to an "lvalue".

In C-land (and in the compiler) the name of this variable is the string "x". In C-land this is often called the identifier of this variable.

In Python you also have variables and identifiers. Example Python REPL (bash also has this):

  >>> x = 5
  >>> x
  5

In Common Lisp they are called symbols instead of identifiers. Think of Python 3's object.__dict__("x") .

Lisp symbols (a.k.a. identifiers a.k.a. variable names) are more powerful and more important than in C89 or Python, because there are source code templates. The most important use-case for source code templates are lisp macros (as opposed to C89 #define-style macros). This is also where backquote and quasiquote enter the picture.

In Lisp you can create a variable name (a.k.a. an identifier a.k.a. a symbol) with the function INTERN (bear with me.)

  (intern "x")

is a bit like adding "x" to object.__dict__ in Python.

Now for QUOTE:

Lisp exposes many parts of the compiler and interpreter (lisp-name "evaluator") that are hidden in other languages like C89 and Python.

Like with the Python example ">>>" above:

We get the string "x" from the command line. It is parsed (leave out a step for simplicity) and the interpreter it told to look up variable "x", gets the value 5 and prints that.

(QUOTE …) means that the interpreter is supposed to give you back the piece of code you gave it instead of interpreting it. So (QUOTE x) or 'x — note the dangling single quote — returns or prints the variable name of the variable named "x".

Better example:

  (+ 1 2)

Evaluates to number 3.

  (quote (+ 1 2))

and

  '(+ 1 2)

both evaluate to the internal source code of "add 1 and 2" one step short of actually adding them.

In source code templates you sometimes provide code that has to be evaluated multiple times, like the iconic increment "i++" has to be evaluated multiple times in many C89 loops. This is where QUOTE is actually useful. (Ignored a boatload of detail for the sake of understandability.)

[panzagl]

First person to ask for more parenthesis in Lisp.

[DadBase]

First time I saw (+ 1 2), I thought it was a typo. Spent an hour trying to “fix” it into (1 + 2). My professor let me. Then he pointed at the REPL and said, “That’s not math—it’s music.” Never forgot that. The '? That’s the silent note.

[no_wizard]

It's due to Polish notation[0] as far as I understand it. This is how that notation for mathematics works.

I suppose my suggestion would break those semantics.

[0]: https://en.wikipedia.org/wiki/Polish_notation

[DadBase]

Aye, Polish notation sure. But what he gave me wasn’t a lecture, it was a spell.

Syntax mattered less than rhythm. Parens weren’t fences, they were measures. The REPL didn’t care if I understood. It played anyway.

[obi2kenobi]

Beautiful. I wish I had more professors who expressed concepts poetically back when I was in school. That's the kind of line that sticks in your head.

[wdkrnls]

R works exactly as you describe. You can type `+`(1, 2) and get 3 because in R everything that happens is a function call even if a few binary functions get special sugar so you can type 1 + 2 for them as well. The user can of course make their own of these if they wrap them in precents. For example: `%plus%` = function(a, b) { `+`(a, b)}. A few computer algebra systems languages provide even more expressivity like yacas and fricas. The later even has a type system.

[seanw444]

Similar in Nim as well.

[ckmate-king-2]

In Standard ML, you can do either 1+2 or op+(1, 2).

[int_19h]

R is much crazier than that because even things like assignment, curly braces, and even function definitions themselves are all function calls. It's literally the only primitive in the language, to which everything else ultimately desugars.

[Reefersleep]

It's not implicit in this case, it's explicit. + is the function you're calling. And there's power in having mathematical operations be functions that you can manipulate and compose like all other functions, instead of some special case of infix implicit (to me, yeah) function calling, like 1 + 2, where it's no longer similar to other functions.

[eyelidlessness]

How is it implicit? The open parenthesis is before the function name rather than after, but the function isn’t called without both parentheses.

If you want to use commas, you can in Lisp dialects I’m familiar with—they’re optional because they’re treated as whitespace, but nothing is stopping you if you find them more readable!

[apgwoz]

, is typically “unquote.” Clojure is the only “mainstream” Lisp that allows , as whitespace. Has meaning in CL and Scheme.

[kazinator]

Ancient LISP (caps deliberate) in fact had optional commas that were treated as whitespace. You can see this in the Lisp 1 Programmer's Manual (dated 1960).

This practice quickly disappeared though. (I don't have an exact time line about this.)

[eyelidlessness]

My mistake!

That’s what I get for not double checking… well… basically anything I think during my first cup of coffee.

[bcrosby95]

func(a, b) is basically the same as (func a b). You're just moving the parens around. '+' is extra 'strange' because in most languages it isn't used like other functions: imagine if you had to write +(1, 2) in every C-like.

[zelphirkalt]

Surely you mean (+ (, 1 2))

;)

[owlstuffing]

> No types barking

No thanks

[tmtvl]

The reason I switched from Scheme to Common Lisp was because I could say...

  (defun foo (x)
    (declare (type (Integer 0 100) x))
    (* x
       (get-some-value-from-somewhere-else)))

And then do a (describe 'foo) in the REPL to get Lisp to tell me that it wants an integer from 0 to 100.

[codr7]

Common Lisp supports gradual typing and will (from my experience) do a much better job of analyzing code and pointing out errors than your typical scripting language.