This post misses the IMO best indentation scheme for lisp, which I used for my college class where we had to use MIT scheme:
(define (match:element variable restrictions)
(define (ok? datum)
(every
(lambda (restriction)
(restriction datum)
)
restrictions
)
)
(define (element-match data dictionary succeed)
(and
(ok? data)
(let ((vcell (match:lookup variable dictionary)))
(if vcell
(and
(equal? (match:value vcell) data)
(succeed dictionary)
)
(succeed (match:bind variable data dictionary))
)
)
)
)
element-match
)
> This post misses the IMO best indentation scheme for lisp [...]
Likewise, a lispier(?) and more compact compromise?:
(define (match:element variable restrictions)
(define (ok? datum)
(every
(lambda (restriction)
(restriction datum))
restrictions))
(define (element-match data dictionary succeed)
(and
(ok? data)
(let ((vcell (match:lookup variable dictionary)))
(if vcell
(and
(equal? (match:value vcell) data)
(succeed dictionary))
(succeed (match:bind variable data dictionary))))))
element-match)It becomes much harder to notice if you have closed all the right matching parens, and hence where a token after the closing paren belongs. When the closing paren is exactly below the opening paren, on the same column, it becomes much easier to see the structure. It's a very good thing for students who just learn Lisp.
But all the cool kids are using paredit or parinfer which means you don't actually ever write the closing parens yourself.
Counting parens is silly when the editor can do it for you.
Likewise, it's better to optimize for experienced readers than it is to optimize for students.
Some time ago parinfer (https://news.ycombinator.com/item?id=42766205) took some attention as a more pythonesque approach to matching presentation and meaning. I liked it, but it seems that it didn’t catch on, alas.
How do you know it didn't catch on? It depends on Lisp! The fraction of programmers using parinfer is bounded by the fraction of Lisp programmers. (I think I might have seen some fledgling experiment with parinfer for other languages, though.)
I'm highly resistant to novelty in this area, yet even I am using parinfer. A crappy implementation in Vim, at that, which only works in One direction: indent -> parens, and can get into an infinite loop fighting with Vim's undo.
Well from what I see on the emacs corners of the internet, not so much discussion.
Also I recently read on an HN thread I started: it was taken out of Calva due to bugs no one fixed.
Not your fault, but I think it's such a silly argument.
About 90% of my problems arise when I shift the parens, e.g. going from `(+ (1 2))` to `(+ (+ 1 2) 3)` or the other way round - and if there is now ) or )) at the very end (but more likely at 4+, not 1 or 2.
How would any plugin know what I mean here? (I know + is a bad example as its commutative...).
Paredit has lots of great ways to do that sort of thing, though, while keeping things balanced. In emacs with paredit, providing a prefix argument to open paren lets you surround n sexps in the new parens. There's also slurp and barf to move sexps in and out of parens.
Editing for lisp mirrors the language as a whole in that there's a learning curve, but once you've climbed it, it's a joy to work with. Because the syntax is so uniform, all you need is a handful of extensions/commands for full blown structural editing.
I get by splendidly in emacs with nothing but paredit and convenient keybindings(not a huge fan of paredit defaults).
You need to clarify your example, because it appears that you are inserting `+` and `3` between existing parens rather than shifting them.
But regardless, I think the answer you are looking for is:
Some of the coolest kids of all use rainbow delimiters so that they can easily identify the matching closing paren with their eyes!
I tried, I'd need to crank the font size up to 14+ to discern the difference ;)
Writing code is usually easier than reading!
In Rust I came to the opposite conclusion.
One webserver framework, Axum, has the trait IntoResponse which is returned by something that implements Handler. A lot of things implement IntoResponse and/or Handler. In plain English you have handlers, these are parts of code that get selectively invoked for a request. The handlers return data that can be made into responses. It's powerful, but if something doesn't fit exactly, the Rust compiler throws gibberish errors at you. It's finnicky as hell.
But once you have written something that works, it is beautiful to read.
So it's difficult to write such code, and if it's easy for you then it's because you mastered the art.
If you're not used to the language, the conceptual models of the codebase you are working on, or you're working with obfuscated code. Otherwise it should be much faster to read and understand the code somebody else wrote than to write it from scratch?
No, I mean, in general, including languages which I work with for 20-25 years. Reading code is still harder than writing.
I can certainly think of examples where reading code is harder than writing it, but given that much of my professional role has been dissecting the horrible code written by other people in order to figure out not only what it does, but also what it was supposed to do, I simply disagree. Reading code is just more boring than writing it.
Exactly.
For Lisp users this is 35% extra vertical space for zero readability benefit. Or negative readability rather. When I see this formatting style I have to fix it before I can read it. It's like trying to read C code with an extra blank line after every single closing brace.
Lisp users are a diverse crowd. I won't be so bold as to sign off for all of them.
The main reason to use it is not for reading it is to work with version control which is all line orientated.
The reading is not that bad as the indentation tells you where the brackets are and you don't really notice them.
Why 35%? I can't seem to align with you on the math here.
You are right, I made the classic mistake of doing the percentage backwards. Once properly formatted I am counting 16 lines and it goes up to 25 so it's a 56% increase.
As a hobbyist that keeps trying to break into lisp, I think the absence of an obviously good formatter for lisps has made things much harder.
This looks like how I've naively tried to format things, and having started with python kind of makes sense.
But trying to read / edit code with 5 parentheses bunched together, I have the hardest time telling if they are matched or where. Thank goodness for vim / helix auto-matching.
The way to have a good time with lisp is using Emacs and Slime/Sly. And if you like shortcuts, you can try Paredit for structural editing.
I just use the Tab key for indenting and it always work for me (there's a logic on where to break line to get correct indenting, and it makes it easy to see if you misplaced a parenthesis. Also, both Emacs and Vim, highlight the matched parenthesis, so I never wonder about if I have the correct number of closing parenthesis.
You can also try "rainbow parens" color scheme that makes nesting much easier to see.
You are looking at the wrong thing. Nobody looks at bunched parentheses for any reason, unless they are investigating a suspicion that the code's indentation is wrong. In that situation, all that matters is whether the bunch contains the correct number of parentheses, not which specific closers match which specific openers.
(foo
(bar
(baz (xyzzy arg))
(bar arg))
(bar
(baz (xyzzy arg))
If this problem is not accidentally compensated by a ) in the wrong place elsewhere, this code will error out.
One simple thing you can do is just automatically reindent the whole block. The indentation will then match the incorrect parentheses and the problem will be clear. Undo the temporary reformatting and fix it.
> One simple thing you can do is just automatically reindent the whole block.
This is exactly the point of my comment -- the absence of an obvious / reasonable automatic formatter is the problem I'm pointing to.
In Vim, use V to go into line-oriented visual selection mode. Move the cursor over the range of lines to be reindented and hit =.
Admit you posted this only to trigger lispers.
I'm a Lisper myself, lol.
That is blasphemy.
You may not like it, but this is what peak lisp indentation looks like :)
Honestly though, this style for ending parenthesis of block-type nodes makes a lot of sense to me. And you still retain the benefit of macro-ing your own syntax when you need it.
The uncuddled closing parens are a sure sign of a sick mind; probably corrupted by C or the like.
They make line-by-line diffs nicer. And it's quite nice to be able to swap/move lines without messing up the closing paren, especially when not using a structural editor.
To be fair in C they don't put semicolon on its own line either.
Maybe they should?
I mean otherwise why not do
void MyFunc() {
DoStuff();
DoOtherStuff();}
Nicely indented easy to read blasphemy is the most dangerous kind!
One thing I dislike about the compact style is how adding/removing a let block causes a diff line that is just adding/removing a parent.
Do you really think it's just a matter of style whether or not you must add or remove a parenthesis? Or whether that change shows up in a diff?
A blasphemy is a very useful thing in cases when a discipline turns into a religion. (Yes, I know about jokes.)
Exactly what I was thinking.
That is
More of an allman Lisp guy myself
(
define (match:element variable restrictions)
(
define (ok? datum)
(
every
(
lambda (restriction)
(restriction datum)
)
restrictions
)
)
(
define (element-match data dictionary succeed)
(
and
(ok? data)
(
let ((vcell (match:lookup variable dictionary)))
(
if vcell
(
and
(equal? (match:value vcell) data)
(succeed dictionary)
)
(succeed (match:bind variable data dictionary))
)
)
)
)
element-match
)I’ve been in too many arguments about Lisp indentation but that actually comes close to decent with a couple of changes: move the singleton closing parens up, and reduce to, say, 2 spaces.
I am sure there are some Emacs modes or plugins for other editors that could go back and forth between this and the more commonly-seen style.
I don't use Lisp but this is very readable to me, actually way more than the examples in the article by a lot. I actually understand this code, just because of how it's formatted. Maybe I'd just use 2 spaces instead of 4 but otherwise, looks good.
Granted I'm not the one to optimize for since I don't use Lisp, but this just seems sane and easy to automate, and it looks like other non-Lisp languages.
Those poor lonely parens!
Do the poor lonely braces in Algol-style languages indented OTBS-style bother you?
Should we be doing this?:
void MyFunc() {
DoStuff();
DoOtherStuff();}Parens are like penguins, they need to snuggle. Get out of here with your vile curly braces.
The only thing that's missing is moving most of nested logic into `let` assignments.
Yeah, I'm surprised the article doesn't address this obvious case. It's pretty much my default indentation scheme for any language, and while I see a lot of lispers here complain about the wasted vertical space, it's undeniably easy to see which parentheses match and what the block structure is.
I'll happily tolerate some wasted space in that trade-off.
Honestly that's the indentation format I use for every language. Doing it in Lisp just makes sense.
My "OTBS all the things" rules:
- Never horizontal align. Indentation is a simple tree, with each set of children indented 1 layer deeper, which can be 4 spaces or 1 tab.
ThisKindOfThing(
foo,
bar,
baz);
- If a block is worth indenting for, the end of the block deserves its own line. Just do OTBS but also for parentheses if you've a long function call.
ThisKindOfThing (
foo,
bar,
baz
);
- if you're ever splitting something across multiple lines there must be an indent.
myObj
.Foo()
.Bar()
.Baz();
Foo(Bar(Baz(
someVal
)));
foo
+ bar
+ baz
+ (
var1
* var2
* var3
)
+ quux;This is exactly the set of rules I happen to have ended up using. I suspect this is one of the few styles that you inevitably arrive at if you search the space for a format that looks okay, is easy to read, doesn't cause things to shift around when functions are renamed, and is relatively simple and self-consistent.
This is mostly how I indent my code. I don't know why so many people hate it. We have huge screens and spacing conveys structure so I use spacing when appropriate, like a ')' in its own line. I work in a *very* small team though and I write most of the code.
That's pretty awful even for other languages.
Is this better?
void MyFunc() {
DoStuff();
DoOtherStuff();}Once you have even a small amount of that background it's complete shit. Or else if it isn't, nobody wants to work with you. But at that point that probably suits you fine. If you want to work completely alone, yet be maximally productive, you've landed in a good language family.
Haha, good one!
(I hope?!)
This is one of the most disgusting things I read on HN.
I love a Lisp-themed article, but this is such a yawn in 2025.
It seems pretty clear that a threading macro which allows the user to place the "hole" being filled wherever they want is the answer.
Example: cl-arrows
(tree-transform-if predicate
transformer
(second tree)
depth)
> Nineteen! Nineteen spaces of indentation! It's getting unruly.
Why, is there a shortage of whitespace these days??
> Such an indent, when used in deeply nested code, makes it too wide and unreadable.
Ah, I see! Well, I would argue the problem is not the indentation style. What makes your code unreadable is that it's too deply nested, and I would bet that no indentation style can help with that.
That's a point too!
Are there any formatters that detect related things and put them in the same column? Here's a Pie example:
(claim Tuple3 (-> U U U U))
(define Tuple3 (lambda ( a b c )
(Pair a (Pair b c) )))
(claim tuple3Of (Pi ( (A U) (B U) (C U))
(-> A B C
(Tuple3 A B C ))))
(define tuple3Of (lambda ( A B C )
(lambda ( a b c )
(the (Tuple3 A B C )
(cons a (cons b c)) ))))
(claim Tuple5 (-> U U U U U U))
(define Tuple5 (lambda ( a b c d e )
(Pair a (Pair b (Pair c (Pair d e))))))
(claim tuple5Of (Pi ((A U) (B U) (C U) (D U) (E U))
(-> A B C D E
(Tuple5 A B C D E ))))
(define tuple5Of (lambda ( A B C D E )
(lambda ( a b c d e )
(the (Tuple5 A B C D E )
(cons a (cons b (cons c (cons d e))))))))> Common Lisp in particular is extremely unfriendly to threading macros. Arrows imply a consistent thread-first or thread-last functions. But CL's standard lib is too inconsistent for that to work. So we're left with picking an indentation style we don't necessarily like.
Diamond arrows exist, you know.
(-<> foo
bar
(baz)
(quux 1 2 3)
(fred 4 5 6 <>)
(frob 7 <> 8 <> 9))
Well, it's all based on my experience. In two Lisp companies I worked in, colleagues complained or even were trying to fix how I indent things. So it's not just some anxiety whatever, it's experience that needs explanation and justification.
I'm aware of arrow-macros/diamond wands, and am using them myself. But I often have to write dependency-less libraries for moderately complicated algos, and that's where indentation style is the only thing I can rely on.
> (tree-transform-if predicate
> transformer
> (second tree)
> depth)
I think this is more a problem with languages that encourage (or don't discourage) deeply nested function calls. I face the same issue in Python, although Python kinda discourages me to do:
foo(
bar(
baz(
far(
faz(...)))))
a = faz(...)
b = far(a)
...
The second Python method that uses intermediary variable is what I use because it works much better with a debugger
I also like this trick to avoid nested for loops:
import itertools
names = ['Bob', 'Sam']
nums = [1,2]
letters = ['A','B']
for name,num,letter in itertools.product(names,nums,letters):
print(f"{name} {num} {letter}")
Bob 1 A
Bob 1 B
Bob 2 A
Bob 2 B
Sam 1 A
Sam 1 B
Sam 2 A
Sam 2 BYou can have the same step-by-step in Lisp via LET*. The problem outlined by the author happens when you have a function call where 1) the operator has a long name, 2) there are many arguments. The TREE-TRANSFORM-IF is such a case and you can't solve it by just "going step by step".
Yes, I think for whatever reason, Lisp languages tend to use long names for funcs and vars whereas Haskell and other FP languages prefer symbols like >>=, <$>, etc. Personally, I prefer somewhere in the middle.
> The TREE-TRANSFORM-IF is such a case and you can't solve it by just "going step by step".
Yes, this is a good point.
Why not just locally bind it to something shorter with let? Lisp is flexible enough that this is a non issue.
You introduce four new variables. You get five lines instead of four, and four of these lines are pure cognitive overhead.
That deep nest of function calls is why I love languages that have bash-style pipelining as a built in feature, either with explicit shell piping features, or where everything Function(myObj) is also implicitly myObj.Method()
Then it becomes (eg PowerShell, a language I hate but use constantly)
Faz |
Far |
Baz |
Bar |
Foo
Faz()
.Far()
.Baz()
.Bar()
.Foo()
I don't know why we're so stuck on prefix notation for function calls. Postfix is just easier to read in long chains.
In Python you have the wonderful https://github.com/pytoolz/toolz
from toolz import pipe
pipe('hello',lambda w: ( "foo", w ), lambda y: [ 0, y, 1 ])
[0, ('foo', 'hello'), 1]
def h(x):
return ( "foo", x )
def g(x):
return [ 0, x, 1]
pipe('hello', h, g)thanks, I'll check this out. I've seen similar libs before, like "pipes" which also does piping through the "|" symbol.
This is a good style indeed, and I'm using it a lot in JavaScript, but you have to admit that it's never as consistent as that—there are basic functions mixed into it, and some lambdas, and whatever.
I prefer a third approach - chaining method calls via `.method()`. C#'s LINQ extension methods and Rust's iterator methods do it and I find it to be the best of both worlds most of the time.
So the "fluent design" approach? I like that too, and many OO languages allow for that, but I haven't seen that approach in functional programming.
Really?
Iterator methods are the only part of Rust/C# that feel like normal decent function composition. Except Haskell is backwards.
It's clear that the blogger is a wide-eyed newbie, who is enamored of Lisp and trying to attract any kind of attention and clicks to it. Obviously, it worked; he got on the front page of HN.
I predict that in another five years, he will write an article disparaging Lisp and purge it from his CV, and that will be that.
I literally program in Lisp for five years now, and I ain't going anywhere.
I just skipped to the bottom of this crap to find the real gem: it's produced with the help of ed.
Yeah, someone working with ed its not going to like 19 repeated leading spaces across lines for vertical alignment. They will like that less than even somebody working with Notepad.
...the author is using ed scripts as a static site generator, in a way completely unrelated to the post content. People are allowed to do oddball things.
Oddball things lead to oddball opinions. Like: you have to go to the monstrous effort of repeating nineteen spaces into the next line again and again to achieve this aligned appearance, therefore it looks ugly.
That's one way to think of it.
> Once you get used to Lisp, you stop noticing the parentheses and rely on indentation instead.
Who does that? LISP has a datastructured syntax, textformat has no meaning in it.
Humans read code. I'd like to become a machine someday, but alas, I still am human and I need to read mine/others' code. Lisp code too.
idk i always just put a newline before every single open parens and that worked for me
never had any real issue using emacs indentation engine (not even sure lisp modes use smie..)
Surely one could do it with different background colors instead. Any examples of that?
For matching parentheses, Emacs users can use a package called rainbow-delimiters that colors matching delimiters. Different levels of enclosure use different colors. I've found this to be too much visual noise -- I am usually just interested on the delimiters for the current s-expressions. In this vein, having multiple different colored backgrounds would be garish and distracting for me.
I have seen it done with foreground colours: https://github.com/alphapapa/prism.el
I think I once saw a project turning Scheme into a visual block-based representation, but I can't find it again.
[dead]
Lisp is an objectively goofy language,* I appreciate someone pointing out the pain points around trying to cobble together readability via indentation.
* Not a bad language - well, maybe bad by today's standards, but undeniably a very important language - just goofy
What languages do you use are readable without indentation? I guess other than Python, where indentation actually contributes towards correctness rather than it just being optional.
Well, there is also https://nim-lang.org as well as a pretty big list at https://en.wikipedia.org/wiki/Off-side_rule . Just expanding, not trying to invalidate your valid question!
APL?
(For certain values of readable)
But well, it's actually the most readable and meaningful language... to APL programmers.