The threading story here is what grabbed my attention. Pass-by-value with copy-on-write means you get data-race immunity without any locks or channels. You just pass data to a thread and mutations stay local. That's a genuinely useful property.

I've worked on systems where we spent more time reasoning about shared state than writing actual logic. The typical answer is "just make everything immutable" but then you lose convenient imperative syntax. This sits in an interesting middle ground.

Curious about performance in practice. Copy-on-write is great until you hit a hot path that triggers lots of copies. Have you benchmarked any real workloads?

Hmm this is a bit like peeling a banana only to throw the banana and eat the peel. Pass by value reduces the true benefit of copy-on-write.

Use immutable pass by reference. Make a copy only if mutability is requested in the thread. This makes concurrent reads lock-free but also cuts down on memory allocations.

I think that what you are calling "immutable pass by reference" is what the OP is calling "pass by value". See, when used abstractly, "pass by value" means that the argument is passed as a value, hence it is immutable and the callee can't mutate it. One way to implement this is by copying the data that represents the value. In the OP's language, and in many other languages that work this way, instead of copying the data, we implement "pass by value" by incrementing the reference count and passing a pointer to the original data. These differing implementations provide the same abstract semantics, but differ in performance.

> Use immutable pass by reference. Make a copy only if mutability is requested in the thread.

This is essentially what Herd does. It's only semantically a pass by value, but the same reference counting optimizations still apply.

In fact, Herd's approach is a bit more powerful than this because (in theory) it can remove the copy entirely if the caller doesn't use the old value any more after creating the thread. In practice, my optimizations aren't perfect and the language won't always detect this.

The big downside is that we have to use atomic reference counts for _everything_. From memory this was about a 5-15% performance hit versus non-atomic counters, though the number might be higher if other bottlenecks were removed.

Peels are rich in fiber.

> Have you benchmarked any real workloads?

Nothing "real", just the synthetic benchmarks in the ./benchmarks dir.

Unnecessary copies are definitely a risk, and there's certain code patterns that are much harder for my interpreter to detect and remove. E.g. the nbodies has lots of modifications to dicts stored in arrays, and is also the only benchmark that loses to Python.

The other big performance limitation with my implementation is just the cost of atomic reference counting, and that's the main tradeoff versus deep cloning to pass between threads. There would definitely be room to improve this with better reference counting optimizations though.

There is some prior work on mitigating the performance cost of immutability that you might be interested in. For example, Clojure's persistent vectors allow fast modifications without destroying the original vector, because internally they're wide trees rather than just linear arrays of memory. This allows for assignments to be implemented without a copy of the full vector. https://hypirion.com/musings/understanding-persistent-vector...

Why don’t we just do this by default for threading in most languages? It’s pretty rare for me to actually want to do memory sharing while threading (mostly because of the complexity)

Because it's super slow and shared memory is super fast. And people generally prefer fast code rather than safe code.

It's not "super slow" and most languages do something very similar within concurrent data structures.

Also, copy by value in itself is just a semantic requirement, it doesn't say how it's implemented.

And shared mutable memory is pretty damn slow (given you are not fine with data race garbage), because atomic operations destroy caches. So it's the usual space-time tradeoff at the end of the day.

Why exactly is imperative syntax "convenient" specifically in the context of inter-thread communication?

He's likely referencing that you would need to use different syntax and style, like re-assigning a variable or chaining calls, like when working with a String in Java.

In C, you can simply mutate the underlying characters. So changing the fifth character in a string is as easy as:

    str[4] = 0;

Whereas using the immutable syntax you might have to do something like:

   str = str.substr(0, 4) + "\0" + str.substr(4);

Well, yes, that's how it becomes convenient in general. But why would you be doing things like that when communicating between threads?

It's only semantically a pass-by-value, in reality a reference is passed and the data is only copied if needed (i.e. value is mutated while shared).

So the language has reference semantics, and (per the edit) for every object (like in Python)?

(Ah, no, your example elsewhere in the thread suggests that the referred-to structures get implicitly copied all over the place.)

Nope, it's value semantics (unlike Python), the references are just an internal optimization. Implicit copies happen when a list/dict with more than one reference is mutated.

> the references are just an internal optimization

Optimization specifically for function calls, or... ?

Because if you're doing all this copy-on-write anyway, the indirection seems to be a needless cost in other contexts.

This applies everywhere, and it fundamentally wouldn't be possible for just function calls.

> needless cost

Are you comparing to a language with mutable references or a our functional language? A language with mutable references will of course be faster, but this is more intended as an alternative to pure functional languages (since functions are referentially transparent).

In this case, the cost of the indirection is approximately zero (relative to the cost of just doing reference counting), since passing a reference just requires a bump to the refcount. And most of the time the refcount increments are skipped by "moving" instead of copying the reference.

I'm comparing to the same language implemented without the supposed internal optimization, according to my understanding of what you're doing.

But I think at this point I'd have to read and analyze the code to have a proper understanding.

... Although granting that you already have paid the cost of reference-counting GC (and, I assume, per-object allocation), it probably is indeed insignificant. And special-casing where the reference count == 1 is also kinda neat. (E.g. CPython doesn't "move references" per se if I'm thinking clearly; but it does detect cases where it can safely mutate the underlying implementation of a type that's "immutable" from the perspective of language syntax.)

(Edit: in the old post title:) "everything is a value" is not very informative. That's true of most languages nowadays. Maybe "exclusively call-by-value" or "without reference types."

I've only read the first couple paragraphs so far but the idea reminds me of a shareware language I tinkered with years ago in my youth, though I never wrote anything of substance: Euphoria (though nowadays it looks like there's an OpenEuphoria). It had only two fundamental types. (1) The atom: a possibly floating point number, and (2) the sequence: a list of zero or more atoms and sequences. Strings in particular are just sequences of codepoint atoms.

It had a notion of "type"s which were functions that returned a boolean 1 only if given a valid value for the type being defined. I presume it used byte packing and copy-on-write or whatever for its speed boasts.

https://openeuphoria.org/ - https://rapideuphoria.com/

> It had a notion of "type"s which were functions that returned a boolean 1 only if given a valid value for the type being defined.

I've got a hobby language that combines this with compile time code execution to get static typing - or I should say that's the plan, it's really just a tokenizer and half of a parser at the moment - I should get back to it.

The cool side effect of this is that properly validating dynamic values at runtime is just as ergonomic as casting - you just call the type function on the value at runtime.

Thanks, I updated the post title based on this and another comment. Thanks for the pointer to Euphoria too, looks like an interesting language with a lot of similar ideas.

I have implemented similar behavior in some of my projects. For one, I also have also implemented 'cursors' that point to some part of a value bound to a variable and allow you to change that part of the value of the variable. I have used this to implement program transformations on abstract parse (syntax) trees [1]. I also have implemented a dictionary based on a tree where only part of the tree is modified that needs to be modified [2]. I have also started working on a language that is based on this, but also attempts to add references with defined behavior [3].

[1] https://github.com/FransFaase/IParse/?tab=readme-ov-file#mar...

[2] https://www.iwriteiam.nl/D1801.html#7

[3] https://github.com/FransFaase/DataLang

At the risk of telling you what you already know and/or did not mean to say: not everything can be a value. If everything is a value then no computation (reduction) is possible. Why? Because computation stops at values. This is traditional programming language/lambda calculus nomenclature and dogma. See Plotkin's classic work on PCF (~ 1975) for instance; Winskel's semantics text (~ 1990) is more approachable.

Things of course become a lot more fun with concurrency.

Now if you want a language where all the data thingies are immutable values and effects are somewhat tamed but types aren't too fancy etc. try looking at Milner's classic Standard ML (late 1970s, effectively frozen in 1997). It has all you dream of and more.

In any case keep having fun and don't get too bogged in syntax.

IMHO this is both unnecessarily pedantic and not really quite right. Let’s say we accept the premise that “everything is a value” means reduction is impossible. But a value is just the result of reducing a term until it is irreducible (a normal form). So if there is no reduction there can’t really be values either—there is just “prose” (syntax) and you might as well read a book.

I am unable to extract any meaning from your post. You appear to be making a general claim: it is impossible to design a programming language where everything is a value. You at least admit that "data thingies" can be values. Are you claiming that it is not possible for functions to be values? (If we assume that the argument and the result of a function call is a value, then this would mean higher order functions are impossible, for example.) If not that, then what? Please give a specific example of something that can never be a value in any programming language that I care to design.

I think parent means it from a lambda calculus perspective. If you only have values at an AST level, then you only have a tree of.. values, like an XML document.

You can apply meaning to a particular shape of that tree which could be executed, but then you basically just added another layer before you parse your AST that becomes executable.

Thanks, some interesting reading there that I will check out (I wasn't aware of PCF). Perhaps I should've used more precise wording: "All types are value types".

> Standard ML [...] It has all you dream of and more

The main thing here that's missing in Standard ML (and most other functional languages) is the "mutable" part of "mutable value semantics" - i.e., the ability to modify variables in-place (even nested parts of complex structures) without affecting copies. This is different from "shadowing" a binding with a different value, since it works in loops etc.

Quick note then a more wordy response (and after being dinged in another thread yesterday, the tl;dr is your usage is correct, ignore purposely avoiding context criticism of wording):

SML has mutation, but only for Ref Cells, which humorously are values themselves. Not that’s what you’re really talking about here.

Now for the wordy part…

As another of your sibling commenters said GP was being incredibly pedantic. While his reference to Plotkin/Winskel and the PCF thread of research is formative, it is of particular note for Structural Operational Semantics.

The real issue GP is raising that in programming language semantics there are two distinct ways in which the term ‘value’ is used. Worse still is that the terms are not just distinct but are broadly from two very distinct fields in PLT. So, what are the two uses:

  1) when the domain of discourse is Operational Semantics of Programming Languages, a ‘value’ is a term in the formal abstract syntax of the language which is not subject to reduction via any other transition defined over the syntax;

  2) when the domain of discourse is the Evaluation and Default Kinds of arguments passed to functions and the semantic implications of those defaults, a ‘value’ is defined in the negative as those semantic objects which ARE NOT references [1];

which from the definitions alone, it is clear your designation of your language as ‘pass-by-value’ is a distinct thing from GP’s usage of the term.

While GP’s usage of ‘value’ is in line with a long standing tradition, that tradition is firmly within the academic/functional programming language syntax and semantics sphere. Your usage, as is PLAINLY APPARENT from context, is absolutely correct and in line with long standing usage within discussion (and academic work) on imperative (and otherwise non-functional) programming language semantics. So keep phrasing discussions about Herd using the ‘pass-by-value’ and ‘everything is a value’. It’s not only contextually correct and historically justified, it is utterly within the ‘vernacular’ of normal programming language discussions.

One last thought, your language’s adoption of totally defaulting to passing arguments (to functions, threads, basically any control construct), with copy-on-write being an optimization only, should make implementing a non-reified linearity qualifier on types relatively simple to implement, that would address some of the locking anti-optimizations and address some of your static analysis that you mentioned where not working 100%.

———————

1: Reference here include Rust/C++ style lightly abstracted references, nearly zero abstraction pointers, and then the references as discussed when talking about Python, Ruby, JavaScript, C++ smart pointers, Rust’s higher level abstractions over references, etc which are a very abstract concept of reference.

Excuse me if I didn't get it right, but as a practical example, I'd assume that I can rewrite every program into JavaScript using the usual control structures and, beyond that, nothing but string values (which are immutable). Simple arithmetic would already be kind of a chore but can be done. (Input and output already happens only via serialized values (cf also webworkers) so there's that; for convenience use TypedArrays wrapped in classes that shield you from immutability). It is not obvious to me where `a = '[1,2]'; a1 = JSON.stringify( JSON.parse( a ).push( 3 ) ) );` is fundamentally different from just pushing a value to a copy of `a`. Also, you could write `a1 = a.slice(0,-1) + '3]'` which only uses non-mutating stuff under the hood.

This sounds quite similar to pure functional languages like Haskell, where a function call cannot have any side effect.

But those go further in that they don't even have any mutable data. Instead of

    var foo = { a: 1 };
    var bar = foo; // make a copy of foo
    set bar.a = 2; // modify bar (makes a copy)

Haskell has

    foo = Baz { a = 1 }
    bar = foo { a = 2 } // make a modified copy of foo

Personally I think local mutability is quite a useful property, which was part of the inspiration for making this instead of just another pure functional language:

- All functions are still referentially transparent, which means we get all the local reasoning benefits of pure functions. - We can mutate local variables inside loops (instead of just shadowing bindings), which makes certain things a lot easier to write (especially for beginners). - Mutating nested fields is super easy: `set foo.bar[0].baz = 1;` (compare this to the equivalent Haskell).

My hobby language[1] also has no reference semantics, very similar to Herd. I think this is a really interesting point in the design space. A lot of complexity goes away when it's only values, and there are real languages like classic APL that work this way. But there are some serious downsides.

In practice I have found that it's very painful to thread state through your program. I ended up offering global variables, which provide something similar to but worse than generalized reference semantics. My language aims for simplicity so I think this may still be a good tradeoff, but it's tricky to imagine this working well in a larger user codebase.

I like that having only value semantics allows us, internally, to use reference counted immutable objects to cut down on copying; we both pass-by-reference internally and present it as pass-by-value to the programmer. No cycle detection needed because it's not possible to construct cycles. I use an immutable data structures library[2] so that modifications are reasonably efficient. I recommend trying that in Herd; it's almost always better than copy-on-write. Think about the Big-O of modifying a single element in an array, or building up a list by repeatedly appending to it. With pure COW it's hard to have a large array at all--it takes too long to do anything with it!

For the programmer, missing reference semantics can be a negative. Sometimes people want circular linked lists, or to implement custom data structures. It's tough to build new data structures in a language without reference semantics. For the most part, the programmer has to simulate them with arrays. This works for APL because it's an array language, but my BASIC has less of an excuse.

I was able to avoid nearly all reference counting overhead by being single threaded only. My reference counts aren't atomic so I don't pay anything but the inc/dec. For a simple language like TMBASIC this was sensible, but in a language with multithreading that has to pay for atomic refcounts, it's a tough performance pill to swallow. You may want to consider a tracing GC for Herd.

[1] https://tmbasic.com

[2] https://github.com/arximboldi/immer

Syntax comment: in your control structures you use a keyword ("do", "then") to start a block as well as wrapping the block in parentheses. This feels superfluous. I suggest sticking with either keywords or parens to delineate blocks, not both.

You should check out Perceus! https://www.microsoft.com/en-us/research/wp-content/uploads/...

But what if mutation is intended? How to pass a mutable reference into a function, so that it can change the underlying value and the caller can observe these changes? What about concurrent mutable containers?

> How to pass a mutable reference into a function, so that it can change the underlying value and the caller can observe these changes?

Just modify the value inside the function and return it, then assign back. This is what the |= syntax is designed for. It's a bit more verbose than passing mutable references to functions but it's actually functionally equivalent.

Herd has some optimisations so that in many cases this won't even require any copies.

> What about concurrent mutable containers?

I've considered adding these, but right now they don't exist in Herd.

The article mentions shallow copy, but does this create a persistent immutable data structure? Does it modify all nodes up the tree to the root?

Yes, if you modify a nested dict/list entry, all nodes above it will be cloned. Here's an example:

  x = [1, [2]];
  var y = x;
  set y.[0] = 3; // clones the outer array, keeps a reference to inner array
  set y.[1].[0] = 4; // clones the inner array here. Outer array is now exclusive so it doesn't need another clone.

  var z = x;
  set z.[1].[0] = 4; // clones both arrays at once

Curious if erlang/elixir isn’t the same sort of thing? Or am I misunderstanding the semantics of “pass by value”?

Your assumption is somewhat correct, for both Erlang and Elixir, however the phrase under discussion doesn’t mean the same thing for immutable languages. Both are ‘pass-by-value’ but that term is being overloaded in a particular way. As I said in another comment, ‘value’ in the language from TFA means any object that IS NOT a reference. The qualifier that every semantic object is a ‘value’ and that therefore, all arguments to a function call, threads spawn, etc are independent values which are (logically, at least) copied to new values that are then passed to the new context.

However, for Erlang and Elixir ‘pass-by-value’ is otherwise called ‘call-by-value’. In this case, it is a statement that arguments to functions are evaluated before they are passed into the function (often at the call site). This is in opposition to ‘call-by-name/need’ (yes, I know they aren’t the same) which is, for instance, how Haskell does it for sure, and I think Python is actually ‘by-name’ as well.

So, Herd’s usage here is a statement of semantic defaults (and the benefits/drawbacks that follow from those defaults) for arguments to functions, and Elixir’s usage is about the evaluation order of arguments to functions, they really aren’t talking about the same thing.

Interestingly, this is also a pair of separate things, which are both separate from what another commenter was pedantically pointing out elsewhere in the thread. Programming language discussion really does seem to have a mess of terminology to deal with.

Nobody has heard of persistent data structures?!

the pipe-equal operator is pretty neat, don't think I've seen any other language do that.

I wrote the following and then realized maybe it is just a quirk of the example in the reader that the ‘set’/‘=‘ pair comes at the end of the chain. If so, it is just a unique syntax sugar for a function, I don’t think it is, so I’m leaving my comment as I wrote it letting this act as a caveat:

Although I don’t particularly like the ‘|’ to be used for chaining functions, I certainly know that it has been a long term syntax coming from Unix. My only issue with the ‘|=‘ is that it should be unnecessary. The only reason I can see that the special operator is required is that the ‘set’/‘=‘ syntax pair is a non-functional keyword ‘set’ with an (I think) overloaded keyword ‘=‘. If the equal sign was an ordinary function (i.e. a function that take a value, and an identifier, associates the value and the identifier, then returns the new value like the Lisps and derived lands) it could just be used arbitrarily in chains of functions.

Nim has a similar, strong preference for value semantics. However, its dynamic heap types (strings, seqs, tables) are all implemented as wrappers that hide the internal references and behave with value semantics by default, unless explicitly escape hatched. It makes it incredibly easy to manipulate almost any data in a functional, expression-oriented manner, while preserving the speed and efficiency of being backed by a doubling array-list.

> In herd, everything is immutable unless declared with var

So basucally everything is var?

I'm not sure if I understand the question?

There are two ways to define a variable binding:

    x = 1; // declares x as immutable
    var y = 2; // declares y as mutable

The "default" behaviour (if no keyword is used) is to define a new immutable variable.

Not everything in C is pass-by-value. Sure, you can argue that a pointer itself is passed by value, but the data it points to is definitely not.

cool project. can you take the address of a variable in some way? i.e. implement your own pointers if its really really needed?

> can you take the address of a variable in some way?

I intentionally didn't add this, mostly because I wanted to explore how far you can get without it (and keep the language simple). Having a "real" pointer as a first class type wouldn't work though, since it would break a lot of the assumptions I use for optimizations.

I did think about two different versions of this but didn't end up adding either:

- Something like `inout` parameters in Swift, which aren't first class pointers. This is really just an alternate syntax for returning multiple values. - A "ref" type, which is essentially a mutable container for an arbitrary value. Passing the container around would share a reference to the same mutable value. This still wouldn't allow modifying values "outside" of the container though.

you're right, once indirection appears pandoras box opens up. keep it as pass by value only, it makes the language unique. although the desire for pointers will never go away, people have used them for so long now.

I think your statement about familiarity is spot on. One of the ‘promises’ of the early functional language efforts (late 70s through mid 90s) was that the invariants made it conceptually simple to have a magical, super compiler take the code and produce binary executables that would surpass imperative languages, due to the aforementioned Pandora’s box effect of the non-functional constructs. Universal value semantics are similar, especially without leaky abstractions allowing “please just let me crack the box, I’ll be good and won’t abuse the power”. Commitment to the invariants, in this case purely (logically, at least) value semantics across the entire language could produce executables that approach low-level, less restrictive code results with a much, much lower burden on the programmer.