OCaml Syntax Sucks (2016)
xahlee.info
xahlee.info
1. there’s no marker to indicate the end of let scopes
2. functions are bound with the same syntax as constants
He asserts that this is confusing. In practice - for the many issues I have with Ocaml! - neither of these are actual issues, in my experience, once code formatting is applied.
An actual serious problem with Ocaml’s syntax is that matches don’t have a terminator, leading people to mess up nested matches frequently. Pair that with the parser’s poor error reporting/recovery and things can become unpleasant quickly.
- Reason, a different syntactic frontend for regular OCaml: https://reasonml.github.io/
- ReScript, a language with OCaml semantics that compiles into: JS https://rescript-lang.org/ (I suppose it's a reincarnation of js-of-ocaml).
But it's not dead, if you look at their GitHub.[1]
Expect some of them to be published at the end of the year or early january!
ReScript isnt a fork of Reason, jsoo focuses only on compiling all OCaml and allows JS interop
Reason with Melange is a viable option for JS/TS devs.
I've had lots of fun playing with Reason a few years ago. I created an interactive real-time visualization tool, to see a Regexp transform into a NFA to DFA to minimal DFA graph: http://compiler.org/reason-re-nfa/src/index.html It only works for basic regexes though.
There is a reason - pun intended - it didn't take off at all with the community.
Not nothing ;-)
The match terminator end thing made me sad when I first saw this in Ocaml. So many languages (C, bourne shell, etc etc) have this exact same problem and it completely sucks in all of them. It's more debilitating in a functional language specifically because matches are more useful than say C case statements so you want to use them much more extensively.
I frequently want to do a pattern match to unpack something and then a further pattern match to unpack further - a nested match is a very intuitive thing to want. Yes you can normally unnest such a match into more complicated matches at the parent level but this is usually much harder for humans to understand.
...and if you had a marker for ending match scopes you could always just reuse that to end let scopes as well if you wanted to although I've literally never a single time run into that as a practical problem (although I haven't written that much OCaml you'd think if it was a real issue I would have banged into it at least once because I found a fair few sharp edges in my time with the language).
Doesn't Bourne shell syntax have terminator keywords for all control flow structures? https://unix.stackexchange.com/a/256152
Case fallthrough almost always ends in tears.
however, my university has a mandatory class taught in ocaml, which i've ta'd for a few times; this is the _number one_ "the undergrad ta couldn't figure out my syntax error" issue students have
I think you’re being generous. The example the author gave is awful because any language can be made illegible if you cram in complicated expressions with multiple levels of nesting into a single line. I’d say it’s outright flamebait.
Apparently, the author hasn't come around to understanding that functions are just another constant.
> because a 0 arity function without side-effect is just constant.
Very often this leads to typing errors. Otherwise it's caught by autoformat (which everyone should be using). But even without that, this is one pitfall every OCaml developer is aware of.
I'm not arguing the syntax is perfect, but I wouldn't say it has a serious problem because of that. Never seen that being a problem.
Yes, it's clearly one of the syntax weak point. You can surround them with parentheses or begin ... end to make things explicit but the default is definitely error prone.
Xah Lee is not exactly known for moderation. If you’re not familiar, browsing the site may be worth your while. It’s full of inflammatory and sometimes bizarre takes, but I’d be hard pressed to argue that it’s ill informed.
Not inclined to do this though, his site doesn't exactly seem easy to browse either. Maybe a reflection of his bizarre takes?
Do you have some examples of well-informed articles by him? I found one about "syntax vs semantics" that seems as inflammatory and ill-informed as TFA under discussion here. I'm not inclined to give this person much leniency...
1. It’s whitespace insensitive, which means I can code something up really messy and the code formatted will automatically fix it up for me.
2. In general there aren’t a ton of punctuation characters that are very common, which is great for typing ergonomics. Don’t get me wrong, there are still a lot of symbols, but I feel compared to some languages such as Rust, they’re used a lot less.
Beyond the syntax, there are a couple of things I really like about the language itself:
1. Due to the way the language is scoped, whenever you encounter a variable you don’t recognize, you simply have to search in the up direction to find its definition, unless it’s explicitly marked as “rec”. This is helpful if you’re browsing code without any IDE tooling, there’s less guessing involved in finding where things are defined. Downside: if the “open” keyword is used to put all of a module’s values in scope, you’re usually gonna have a bad time.
2. The core language is very simple; in general there are three kinds of things that matter: values, types, and modules. All values have a type, and all values and types are defined in modules.
3. It’s very easy to nest let bindings in order to help localize the scope of intermediate values.
4. It has a very fast compiler with separate compilation. The dev cycle is usually very tight (oftentimes practically instantaneous).
5. Most of the language encourages good practice through sane defaults, but accessing escape hatches to do “dirty” things is very easy to do.
6. The compiler has some restrictions which may seem arcane, such as the value restriction and weak type variables, but they are valuable in preventing you from shooting yourself in the foot, and they enable some other useful features of the language such as local mutation.
I never really seen someone put that into words. I always feel a certain kind of weird when I look at a language with tons of punctuation (Typescript is good example).
Would be curious if there's work into a programming language that was optimized for minimal typing while iterating.
The accent grave (backtick) you're using in a Markdown-inspired way is utterly annoying to type on keyboards where accents are implemented as dead keys, to be combined with vowels, common on european keyboard layouts. For your example I had to type `, but then look ahead to the a of add_numbers and remember to put in an extra space, so that the ` and the "a" don't combine to an "à".
Also I find it somewhat illogical: The usage of an accent as a character in itself in programming languages is one of my pet peeves. Just being an ASCII character is not reason good enough to keep using it.
Curly or square brackets, backslashes and other stuff also put you into uncomfortable AltGr or Cmd-Shift territory on some keyboards. American language designers are often blind for these simple ergonomics.
I understand that keyboards behave as they behave, and saying that "well they shouldn't" isn't a real answer. But I also think that it's not true that the backtick character is an accent grave, even if it is treated like one, any more than any other pair of Unicode homoglyphs are the same.
In other languages, modern IDEs take care of that. For an invocation like `addNumbers(1, 2)`, you would perform code completion after typing “aN”, and the IDE would already fill in matching function arguments from the current scope. The selection (cursor) would be highlighting the first argument, so you can readily replace it. Pressing Tab would move to the next argument. Pressing Enter (or similar) would accept the call expression in the current state and move the cursor behind it. So you only type the actual arguments (if necessary) and otherwise just hit combinations of Tab and Enter, or Ctrl+Space for code completion. You generally don’t type the parentheses or commas yourself.
That being said, I’m in favor of languages with not too much punctuation. But there’s a balance, and too little punctuation can also hurt readability.
(x) => { ... }
Punctuation is used in written human languages to provide assistance to our brain-parsers—like road signs that help navigate a sentence. Too much punctuation and it becomes a huge problem because it ceases to be meaningful, but have you ever tried reading a Roman inscription written with no spaces or sentence boundaries?
I think programming language punctuation serves the same role—it visually marks out how the parser is going to navigate the code, allowing the reader to use less mental effort repursing code that they initially misparsed. ML-style languages have a simpler syntax than the C family, which means there is some justification for having less assistive punctuation, but I've definitely struggled to navigate OCaml as a result of too little punctuation.
On the other hand, I like that there's little overloaded syntax, and the meaning of different characters is fairly consistent.
I've considered trying it for Advent of Code this year.
(also: generally speaking the "core" of .NET is quite unlike most other products Microsoft puts out, so chances are you will be pleasantly surprised)
I'm pretty sure it's "working on a codebase" that kills your soul, not the minutae of particular language choice.
The author doesn't understand why OCaml (and Haskell, and many HM-derived languages) have let ... in ... at all, as opposed to having, say, just lambdas. That's because of a feature of Hindley–Milner type systems called let-polymorphism[1].
This is hard to see in OCaml if you are not intimately accustomed with HM type systems because definitions always use let, but in Haskell it's immediately clear that something more interesting is happening because regular definitions do not use let.
F#, which has been heavily inspired[2] by OCaml got rid of the in, rather it uses indentation to do automatic in insertion[2]. OCaml does not want to be indentation-aware.
[1] https://en.wikipedia.org/wiki/Hindley–Milner_type_system#Let...
If one wanted to criticize OCaml syntax, the need for .mli-files (with different syntax for function signatures) and the rather clunky module/signature syntax would be better candidates.
Sometimes I wrote (haven't written OCaml for some time now..) functions like:
let foo: int -> int = fun x ->
..
val foo: int -> int
"Let x be equal to 4 in x * x" is a natural way to express a concept in English.
The equivalent OCaml code is:
let x = 4 in
x * x
In fact, the `in` makes OCaml's scoping rules some of the easiest to follow. This is valid OCaml code:
let x = 4 in
let y = 3 in
let z = 2 in
x * y * z
let x = 5 in
let y = 6 in
let z = 7 in
x * y * z
I have an easier time understanding scope in OCaml, which explicitly spells out its scoping rules in plain-English syntax, than I do in Rust or JavaScript, which use the deceptively-simple-looking braces but also wild concepts like lifetimes or hoisting.
let x = 1
and y = 2
and z = 3 in
x * y * z
After reading OCaml From The Very Beginning my stance softened. Learning PURE OCaml is a must. (if you want to do AoC with it). You may find yourself bringing in Core or Base, after you've got a good idea of a solution.
It's definitely a fun language for doing AoC... but it can become extraordinarily frustrating when you hit a wall and all the StackOverflow answers assume Core.
The worst thing so far is that you have to declare and define functions before using them. This results in unimportant utility functions being at the top, and the most important functions being at the bottom of a file.
I haven't had much issue with let bindings, however that might be because my functions are fairly simple for now.
That's an extremely good thing.
Unordered definitions and where are some of Haskell worst decisions with lazy by default, overuse of operators and a community which sadly think point-free is desirable.
That's the kind of choices which makes Ocaml a lot more readable.
However, I agree on laziness, operators and point-free.
I've read other articles and there is other weird stuff. Like Java has perfect syntax because on one line you could do so less. In the mean time modern "functional" (or "monadic") style java is a chained mess with ridiculously long lines.
I have seen many devs that prefer the Java imperative style over more functional chained style. Could be an outcome of leetcode style interviews or that most CS programs start with C/python/Java style languages but it's not uncommon to see that preference especially among junior devs
http://xahlee.org/Periodic_dosage_dir/las_vegas/20031015_cop...
Found it via Google
I found F# to be more pleasant to work with async than C# (which is already a breeze). It is true that you still have to define 'task' (or 'async' if you want to use Async CEs) but it is generally there for a reason. I don't think it's too much noise:
let printAfter (s: float<second>) = task {
let time = TimeSpan.FromSeconds (float s)
do! Task.Delay time
printfn $"Hello from F# after {s} seconds"
} let async printAfter (s: float<second>) =
let time = TimeSpan.FromSeconds (float s)
await Task.Delay time
printfn $"Hello from F# after {s} seconds"
'async'-annotated methods in C# enable 'await'ing on task-shaped types. It is bespoke and async-specific. There is nothing wrong with it but it's necessary to acknowledge this limitation.
let!, and!, return!, etc. keywords in F# are generic - you can build your own state machines/coroutines with resumable code, you can author completely custom logic with CEs. I'm not sure what led you to believe the opposite. `await Task.Delay` in C# is `do! Task.Delay` in F#. `let! response = http.SendAsync` is for asynchronous calls that return a value rather than unit.
In the same vein, seq is another CE that is more capable than iterator methods with yield return:
let values = seq {
// yield individual values
for i in 1..10 -> i
// yield a range, merged into the sequence
yield! [11..20] // note the exclamation mark
}
Reference: https://learn.microsoft.com/en-us/dotnet/fsharp/language-ref...
I would disagree. If you need to have a bespoke set of syntax, then something is not integrated where it should be. The language design should not be such that you are writing things differently, depending on the paradigm that you're handling. That's not something that occurs in every language, so it isn't essential that it exists.
We can acknowledge the differences in a way that alerts the programmer, without forcing the programmer to switch syntaxes back and forth when moving between the paradigms. async/await is one method, Promises another, etc. A different syntax is a much, much higher cognitive load.
Support of asynchronous code and of its composition is central to C#, which is why it does it via async/await and Task<T> (and other Task-shaped types). Many other languages considered this important enough to adopt a similar structure to their own rendition of concurrency primitives, inspired by C# either directly or indirectly. Feel free to take issue with the designers of these languages if you have to.
F#, where async originates from, just happens to be more "powerful" as befits an FP language, where resumable code and CEs enable expressing async in a more generalized fashion. I'm not sold on idea that C# needs CEs. It already has sufficient complexity and good balance of expressiveness.
Do-notation-like 'await' is not for calling functions, it is for acting on their return values - to suspend the execution flow until the task completes.
However, the compiler does not, has never, required that it does things via a different syntax. In fact, in the early branches before that was adopted, it didn't! The same behaviour was seen in those branches. This behaviour you expect, was never something that had to be. It was something chosen to simplify the needs of the optimiser, and in fact cut the size of code required in half. It was to reduce the amount of code needed to be maintained by the core team. And so 1087 [1] was accepted.
So perhaps you might need to read more into the process of the why and how async was introduced into C# and F#. It was a maintenance team problem. It was a pragmatic approach for them - not something that was the only way that this became a possibility.
As said, in the original branch for using tasks...
> Having two different but similar ways of creating asynchronous computations would add some cognitive overhead (even now there are times when I am indecisive between using async or mailbox for certain parallelism/concurrency scenarios). [0]
[0] https://github.com/fsharp/fslang-suggestions/issues/581
[1] https://github.com/fsharp/fslang-design/blob/main/FSharp-6.0...
However, this is where our opinions differ. I like task CE (and taskSeq for that matter too). It serves as a good and performant default. It's great to be able to choose the exact behavior of asynchronous code when task CE does not fit.
Simple things like the maybe and either monad are often clearer in this notation. Complex things like alternatives to async (such as CSP derived message passing concurrency), continuations, parser combinators, non-determinism, algebraic effects and dependency tracked incremental computations are naturally modeled with this same machinery, with CE notation being a kind of super helpful DSL builder that makes certain complex computations easier to express in a sequenced manner.
If the custom syntax was only for async you'd have a point, but the general power of the framework make it the more preferable approach by far, in my opinion.
We shoehorn things that feel like, but are structurally different, to DSLs into config files and the like, using JSON/YAML/etc in rough ways, because DSLs introduce a cognitive overhead that doesn't need to be there.
That the shoehorn happens, does mean that DSLs are something natural to reach for. You're right there. But that we have moved away, as an industry, indicates that using any kind of DSL is a smell. That there probably is a better way to do it.
Having a core language feature using a DSL, is a smell. It could be done better.
Rephrased: ocaml is so flexible that async can be implemented as a library with no special support from the language.
This is the beauty of ocaml (and strongly typed functional languages more broadly)
I don't think that's anything specific to strongly typed functional languages. In eg Rust even the standard library relies on third party crates.
Though it is still somewhat amusing to me that loops in Haskell are delivered via a third party library, if you ever actually want them. See https://hackage.haskell.org/package/monad-loops-0.4.3/docs/C...
I do agree that it's good language design, if you can deliver what would be core functionality via a library.
Whether you want to integrate that library into the standard library or not is an independent question of culture and convenience.
(Eg Python does quite well with its batteries-included approach, but if they had better dependency management, using third party libraries wouldn't be so bad. It works well in eg Rust and Haskell.)
Clojure has core.async, which implements "goroutines" without any special support from the language. In fact, the `go` macro[1] is a compiler in disguise: transforming code into SSA form then constructing a state machine to deal with the "yield" points of async code. [2]
core.async runs on both Clojure and ClojureScript (i.e. both JVM and JavaScript). So in some sense, ClojureScript had something like Golang's concurrency well before ES6 was published.
[1] https://github.com/clojure/core.async/blob/master/src/main/c...
[2] https://github.com/clojure/core.async/blob/master/src/main/c...
That's wildly overselling it. Closure core async was completely incapable of doing the one extremely important innovation that made goroutines powerful: blocking.
(let [c (chan)]
;; creates channel that is parked forever
(go
(<! c)))
c := make(chan interface{})
// creates goroutine that is parked forever
go func() {
<-c
}()Call a function that blocks in go, the routine will park. Do that in Clojure and the whole thing stalls.
There is a caveat where Java code using `synchronized` will pin a carrier thread, but this has been addressed in recent versions of Java.[1]
Because JavaScript runs an event loop on a single thread. It's akin to using GOMAXPROCS=1 or -Dclojure.core.async.pool-size=1 . There may be semantic differences depending on the JavaScript engine, but in the case of web browsers, the only function that could possibly stall the event loop is window.alert. As for Node.js, one would have to intentionally use synchronous methods (e.g. readFileSync) instead of the default methods, which use non-blocking I/O.
When using core.async in ClojureScript, one could use the <p! macro to "await" a promise. So there is no use for Node.js synchronous methods when using core.async (or using standard JavaScript with async/await).
I would call this "making use of the platform" rather than "special support from the language". The core.async library does not patch or extend the base language, and the base language does not patch or extend the platform it runs on.
It's early days in that regard, with some folks doing some really interesting things: Odersky himself / the Ox project.
Gleams a nice one too if you want pure functional, small community though.
It's like... when you mismatch brackets or braces in a C-style language, except to resolve the problem you can't just find the bracket that's highlighted in red and count; you have to read an essay.
I don't know why there are so many people here defending it. It's pretty clearly very elegant, but extremely inconvenient.
* An auto-formatter (ocamlformat integration in your editor, or ocaml-top) that shows how the actual nesting looks like
* You can add ;; at the end of a top-level function to get a syntax error at a better location
* Use the LSP integration of your editor which will show you where the error is as you type, so you catch the problem early
Yeah I do have this set up and it's a very good LSP, but unfortunately frequently if you get the nesting wrong the error is like "the entire rest of the file is wrong". It's often not very helpful.
That's not unique to OCaml. C style syntaxes can give "everything after this point is an error" if you get brackets wrong, but it's much easier to figure out what you did wrong.
type 'a Tree =
| Node of ('a Tree, 'a Tree)
| Leaf of 'a
Then when you concrete-ize it, you have an "int Tree" or a "string Tree".
This is a little clearer with thinking of the List datatype: this is a number list, that's a string list, that's a grocery list, this is a guest list...etc.
One nice thing in F# is that there's an ability to "standardize" how you write generics to look more like the (admittedly arbitrary) way that C# or Java or C++ write them:
type Tree<'a> =
| Node of (Tree<'a>, Tree<'a>)
| Leaf of 'aThat's what I was referring to when I said "as a non-native english speaker". In my native language I would say those in a way that's more similar to "a list of numbers", etc etc.
> One nice thing in F# is that there's an ability to "standardize" how you write generics to look more like the (admittedly arbitrary) way that C# or Java or C++ write them:
Arguably this order (type then generic arguments) is not that arbitrary because this is effectively a type level function, so it makes sense that it uses the same order. However the angle brackets are indeed arbitrary, and IMO to be more consistent the same syntax as function application should be used. For practical purposes though this may make a type's syntax too similar to an expression, so it may not be the best choice (unless you're in a language where types are actually values)
I mentioned the typing experience because I hate to have to 1. open and close braces 2. press shift while typing
[2] is also the reason I prefer generics in python and scala than that of java and C#
It's a checkbox at most web hosts, built in to many reverse proxies, etc. There's no excuse for not offering htttps, particularly since it places users at risk if at any point along the path between them and you, there's someone untrustworthy.
Help me out here, what's the threat model here while reading troll programming blogs?
What is the threat model?