I didn't say nullability is orthogonal to value types; I said it was orthogonal to the two-projections world, which is what that text in the article was about rather than nullability.
As to value types and null, I'm not sure about the current picture, but the general idea is that you declare what semantic properties you want - identity or not, nullable or not, tearable or not - and then the compiler picks the best technical in-memory representation for each use. For example, the compiler could choose not to flatten variables that could be null in the heap but to flatten them in the stack. That's the general idea, but I'm not sure about the details, some of which may yet change.
More generally than just Java, nullability is often a property not of a type but of a variable. For example, in C, an int may not be null, but a pointer to an int may be. Now, in C, `int` and `int*` are two different types, but that's exactly a distinction that the original projection-spit design made and we wanted to avoid. But you could still end up with a variable that could hold either an integer or a null and another that may hold an integer but not a null, only this is separate from the reference/value projection, which combines both identity and nullability (in C, `int*` is not only nullable, but also has identity).
> More generally than just Java, nullability is often a property not of a type but of a variable
I'm going to hard disagree here. And the syntax proposed in the Null-Restricted Value Class Types JEP is a major step backwards.
I want to banish nulls from my codebase, completely. I can currently do this with a variety of annotations (at the package-info.java level) and tooling, though it's not integrated well with the language.
Forcing exclamation marks into every variable and parameter is a lot of annoying noise that quite simply nobody will do. The default should be non-nullable, especially for value types.
Declaring whole types as non-nullable is less noisy and errorprone than annotating every variable declaration. If you aren't going to give me "declare the whole codebase as non-nullable" then at least give me something coarse-grained.
> the syntax proposed in the Null-Restricted Value Class Types JEP is a major step backwards. Forcing exclamation marks into every variable and parameter is a lot of annoying noise that quite simply nobody will do. The default should be non-nullable, especially for value types.
Then you didn't read the JEP draft (it's not an accepted JEP) carefully. It says, under "future work":
Providing a mechanism in the language to assert that all types in a certain context are implicitly null-restricted, without requiring the programmer to use explicit ! symbols.
In other words, the draft already incorporates your point, but JEPs (both drafts and actual JEPs) follow the pattern we've found to work so well, that features are best delivered piecemeal rather than in a big bang.
Having said that, I don't know the current plans for this matter, as that document is only in Draft status, so saying it's good or bad is pointless, as it's not even a proposal yet, just something being explored.
It's a proposal for a proposal, fine. Consider this my proposed feedback.
I don't think this syntax is desirable as currently proposed, and that one line under "future work" is doing far too much lifting. My sincere hope is that there are people closer to the process that also feel this way, they will provide similar feedback, and the next draft will be something completely different.
Except we like delivering features by pieces, so unless your proposal is to first deliver global flags with no instance-specific control and later add more control (and I think you're not suggesting that), I don't see any difference between what you're suggesting and what that draft is suggesting. That a one-line under "future work" is doing a lot of lifting is just how we usually do it (and it's okay, people always complain, but we've tried the big-bang approach and this one just works better for us).
So consider that draft as an idea for how the site-specific nullability annotations could work rather than an idea for how nullability could work in the language in general.
> how the site-specific nullability annotations could work
That's fine but it's still a bad choice. The problem is that it defaults to everything nullable and adds noise for non-nullable. This is backwards; in the overwhelming majority of software that people use Java for, non-null is the common case and nullability is the exception.
Kotlin and Typescript got this right. Nobody wants! to! write code! like! this! And they won't, so the feature won't get used and we're back to everything being nullable.
Except there's no choice here. It's a draft of an exploration of a portion of a feature.
Drafts are ideas that have not even been submitted for consideration for inclusion in the roadmap. You can find drafts over ten years old that have long since been superseded by other ones or abandoned altogether: https://openjdk.org/jeps/0#Draft-JEPs. Some JDK engineers write JEP drafts when they feel they get close to something they would like to propose, while others write drafts for pretty much any idea they have.
> The problem is that it defaults to everything nullable and adds noise for non-nullable
It doesn't, though. Even if this draft ever becomes a JEP (and I don't know if it or anything like it will), in its present form or another, it would still, like most JEPs, describe only part of the feature. It's perfectly on point for one JEP to describe the explicit nullability annotations, while another describes the defaults. Smaller features than this have been split into two or three JEPs. This is just how Java features have been described and delivered for years now (see how many JEPs patterns were split into: https://openjdk.org/jeps/0).
It's perfectly okay to dislike some design, but I find it strange to assume, based on a draft of a portion of a feature that one of the most experienced and successful programming language design teams in the history of software is likely to get it wrong. Maybe wait until there's an actual proposal and a roadmap for the complete feature before critiquing it? It's like seeing a draft of some building's foundations by a prestigious architectural firm and saying, these idiots forgot to plan a roof.
This is a strange conversation; you must be misinterpreting me.
I'm not saying that the people who wrote this are dumb, or that Java is a bad language, or that it's time to move to Kotlin, etc. I use Java every day. I pick Java for greenfield projects. I'm on Team Java.
I'm saying I don't like this proposal. I happily accept that this proposal might not become part of the Java language. That's great! I'm spending time writing this in public specifically to discourage it in some tiny way.
It would be fair to tell me "you should write this to the proposal writers directly". And I understand how the language snobbery on HN might make you feel defensive - I often feel that too. But I think my criticism here is valid.
And I'm trying to tell you that this document is not what you think it is. It's a rough sketch of a building's foundations and your critique is about the roof. Even if this were to become a proposal, it's likely the matter of defaults of this feature will be covered by a different JEP, because Java features are usually broken down into multiple JEPs. What you're complaining about may be part of the feature, but it is not covered by this document. If this becomes a JEP and another JEP talks about nullabilty defaults, then you could criticise the selection of defaults, but that particular aspect of this feature is outside the scope of this particular document. So one, this is not a proposal, and two, you haven't seen the description of the part of the feature you want to criticise.
We are well aware that splitting features over multiple JEPs can invite such misunderstandings, but that doesn't change the fact that Java features are, at least currently, split over multiple JEPs. We are also well aware that if this does become a proposal, adding ! everywhere is not what we want, but we want to cover that aspect in a different document, as we usually do. Most Java users aren't confused by this because they don't read the JEPs at all, but such splits help focus the discussion on one aspect of a feature at a time. So your desire for a non-nullable default is very understandable, it's just not relevant as a critique of this document.
For example, the virtual threads JEP described a pinning limitation. We knew it reduced the applicability of that JEP and said as much. We just wanted to address it in a different JEP, and so we did (https://openjdk.org/jeps/491). Ever since the JDK switched to time-boxed, semiannual releases, this is how we've delivered features: in multiple pieces. The same applies to Valhalla.
Yes, but it comes from Java having both runtime and compile-time types; it's harder to make the distinction in languages that don't have runtime types.
In Java, you can ask, `x instanceof T` (and this is a runtime test), which means, is x one of the values in the set of values allowed by T. `null instanceof Integer` is false, even though a variable of type `Integer` can be assigned a null. So you can think of `Integer x` as being `Integer|null x`, i.e. x can hold a null, even though `null instancof Integer` is false.
I think I mostly got this, but just to test it, it would be like in Typescript where I might say:
type Foo = { x: number; }
type Bar = { x: number; y: number }
type FooBar = Foo | Bar;
function baz(x: FooBar) {
if ('y' in x) {
// compiler now knows x is a Bar
}
}
In this case, the variable `x` has a property that is determined by the compiler based on control flow. i.e. it isn't explicitly carried by the type of `x`.
As to value types and null, I'm not sure about the current picture, but the general idea is that you declare what semantic properties you want - identity or not, nullable or not, tearable or not - and then the compiler picks the best technical in-memory representation for each use. For example, the compiler could choose not to flatten variables that could be null in the heap but to flatten them in the stack. That's the general idea, but I'm not sure about the details, some of which may yet change.
More generally than just Java, nullability is often a property not of a type but of a variable. For example, in C, an int may not be null, but a pointer to an int may be. Now, in C, `int` and `int*` are two different types, but that's exactly a distinction that the original projection-spit design made and we wanted to avoid. But you could still end up with a variable that could hold either an integer or a null and another that may hold an integer but not a null, only this is separate from the reference/value projection, which combines both identity and nullability (in C, `int*` is not only nullable, but also has identity).