Great talk about typesystems. It's really nice what you can do with product types, sum types, and exhaustive switching. That's why I like TS. And I'm really looking forward to C# 15 and closed hierarchies (and unions, of course).
The recording of my talk from Software Should Work is now up. :) www.youtube.com/watch?v=0BXu...
The Unreasonable Effectiveness of Constructive Data Modeling - Alexis King | SSW 2026
YouTube video by Software Should Work
www.youtube.com
November 4, 2026 at 9:01 PM
Great talk about typesystems. It's really nice what you can do with product types, sum types, and exhaustive switching. That's why I like TS. And I'm really looking forward to C# 15 and closed hierarchies (and unions, of course).
The upgrade to 1.19 continues and we discover a long-time convenience being snatched away in favour of more explicit syntax as Elixir adopts gradual set-theoretic types.
zarar.dev/struct-updat...
#elixir #elixirlang #typesystems
zarar.dev/struct-updat...
#elixir #elixirlang #typesystems
Struct Updates Now Require Pattern Matching in Elixir 1.19
After upgrading to Elixir 1.19, you might see warnings like this:
The fix is to pattern match on the struct when you define the variable:
Both ...
zarar.dev
January 2, 2026 at 5:38 PM
The upgrade to 1.19 continues and we discover a long-time convenience being snatched away in favour of more explicit syntax as Elixir adopts gradual set-theoretic types.
zarar.dev/struct-updat...
#elixir #elixirlang #typesystems
zarar.dev/struct-updat...
#elixir #elixirlang #typesystems
Ishan Bhanuka showing us "Type System Foundations for Super-Charged Semantic Code Search" #live at #FnConf 2025.
#TypeSystems
#TypeSystems
January 25, 2025 at 10:01 AM
Ishan Bhanuka showing us "Type System Foundations for Super-Charged Semantic Code Search" #live at #FnConf 2025.
#TypeSystems
#TypeSystems
Master bidirectional type inference in Elm. Join our hands-on tutorial at ##FnConf: https://confengine.com/conferences/functional-conf-2025/proposal/21310?utm_source=bluesky&utm_medium=organic_social&utm_campaign=FnConf-speaker-profiles ##TypeSystems
December 20, 2024 at 3:05 PM
Master bidirectional type inference in Elm. Join our hands-on tutorial at ##FnConf: https://confengine.com/conferences/functional-conf-2025/proposal/21310?utm_source=bluesky&utm_medium=organic_social&utm_campaign=FnConf-speaker-profiles ##TypeSystems
I think this is definitely more 'people like to chat about typesystems' than making that much of a material difference for LLM-tooled programming. I doubt the difference between say, a C#-ish and a rust-ish typesystem matters, and I'd semi-weakly guess go/rust doesn't actually matter either.
February 13, 2026 at 9:15 PM
I think this is definitely more 'people like to chat about typesystems' than making that much of a material difference for LLM-tooled programming. I doubt the difference between say, a C#-ish and a rust-ish typesystem matters, and I'd semi-weakly guess go/rust doesn't actually matter either.
Type-constrained code generation could be the next frontier for LLMs. Smaller AI labs might outperform giants by specializing in niche programming languages. Combining general models with fine-tuned specialists could be the winning formula! 🧠💻
#AI #Programming #CodeGeneration #TypeSystems
#AI #Programming #CodeGeneration #TypeSystems
Type-constrained code generation with language models | Hacker News
The article discusses the potential of type-constrained code generation using language models, particularly focusing on the advantages of specializing models for specific programming languages. It highlights the idea that smaller AI labs could excel by creating highly specialized models for less common languages, as larger models may not perform as well in niche areas. The discussion includes insights on how type systems, like those in TypeScript, can enhance the performance of language models in code generation, but also notes the challenges they face with complex types and error handling. A key takeaway is the suggestion that combining a general model with specialized fine-tuned models could yield better results in code generation tasks.
news.ycombinator.com
May 19, 2025 at 12:03 AM
Type-constrained code generation could be the next frontier for LLMs. Smaller AI labs might outperform giants by specializing in niche programming languages. Combining general models with fine-tuned specialists could be the winning formula! 🧠💻
#AI #Programming #CodeGeneration #TypeSystems
#AI #Programming #CodeGeneration #TypeSystems
Typographic Universe
ku-viscom.com
June 5, 2025 at 1:28 PM
ask me about typesystems (don't)
February 24, 2025 at 10:16 PM
ask me about typesystems (don't)
Developer builds feed aggregator in Gleam, shares insights on type-safety and Erlang VM integration
https://olano.dev/blog/gleam-coming-from-erlang
#programminglanguages #gleam #erlang #typesystems #feedaggregator
https://olano.dev/blog/gleam-coming-from-erlang
#programminglanguages #gleam #erlang #typesystems #feedaggregator
February 26, 2025 at 4:17 AM
Developer builds feed aggregator in Gleam, shares insights on type-safety and Erlang VM integration
https://olano.dev/blog/gleam-coming-from-erlang
#programminglanguages #gleam #erlang #typesystems #feedaggregator
https://olano.dev/blog/gleam-coming-from-erlang
#programminglanguages #gleam #erlang #typesystems #feedaggregator
New functional language Neut brings static memory management without GC or type annotations
https://vekatze.github.io/neut/overview.html
#programminglanguages #memorymanagement #typesystems #functionalprogramming #developertools
https://vekatze.github.io/neut/overview.html
#programminglanguages #memorymanagement #typesystems #functionalprogramming #developertools
February 22, 2025 at 1:07 AM
New functional language Neut brings static memory management without GC or type annotations
https://vekatze.github.io/neut/overview.html
#programminglanguages #memorymanagement #typesystems #functionalprogramming #developertools
https://vekatze.github.io/neut/overview.html
#programminglanguages #memorymanagement #typesystems #functionalprogramming #developertools
Java's decision to put equals & hashCode in the Object type was a brilliant resolution to an incredible number of equality &identity defects that still exist in dynamic language typesystems, particularly for collections,don't@ me
(if you dislike Java&want a modern frontend to a great vm, use Scala)
(if you dislike Java&want a modern frontend to a great vm, use Scala)
PHP, this is why no-one likes you:
> json_encode( new Set([new Set([1,2]),new Set([2,1])]) );
= "[[1,2],[2,1]]"
Incorrect, the most popular kind of correct
> json_encode( new Set([new Set([1,2]),new Set([2,1])]) );
= "[[1,2],[2,1]]"
Incorrect, the most popular kind of correct
February 21, 2026 at 8:00 PM
Java's decision to put equals & hashCode in the Object type was a brilliant resolution to an incredible number of equality &identity defects that still exist in dynamic language typesystems, particularly for collections,don't@ me
(if you dislike Java&want a modern frontend to a great vm, use Scala)
(if you dislike Java&want a modern frontend to a great vm, use Scala)
There's a lot affine typesystems can do! The standard one is that you can determine object lifetimes with static analysis, allowing GC to be foregone. Another is a subset of the affine typesystems are the polynomial typesystems, which can prove bounds on program runtime at the type level
June 17, 2025 at 12:51 PM
There's a lot affine typesystems can do! The standard one is that you can determine object lifetimes with static analysis, allowing GC to be foregone. Another is a subset of the affine typesystems are the polynomial typesystems, which can prove bounds on program runtime at the type level
Speaking of typesystems that eliminate entire classes of bugs, have there been any whispers that Haskell might get on the affine bandwagon sometime soon?
June 16, 2025 at 6:34 PM
Speaking of typesystems that eliminate entire classes of bugs, have there been any whispers that Haskell might get on the affine bandwagon sometime soon?
How big is this set? Mot sure, we really just wanted to do a thing and make it work for this paper. But we believe we can create typesystems that restrict the set of programs we might search over in a synthesis such that, taken as a search space, it is polynomial in size
April 24, 2025 at 1:55 AM
How big is this set? Mot sure, we really just wanted to do a thing and make it work for this paper. But we believe we can create typesystems that restrict the set of programs we might search over in a synthesis such that, taken as a search space, it is polynomial in size
In short, there exist typesystems where all programs that do not run in polynomial time are never well-typed. The set of well-typed programs in these typesystems, evidently, is a subset of the set of well-typed programs in an unrestricted typesystem.
April 24, 2025 at 1:55 AM
In short, there exist typesystems where all programs that do not run in polynomial time are never well-typed. The set of well-typed programs in these typesystems, evidently, is a subset of the set of well-typed programs in an unrestricted typesystem.
I'm hung up on the presentation of this one because I don't think I'm gonna do for substructural typesystems what Rust did for affine typesystems; this is a six-pager and that's not the point. But I really need to do better at presenting it than this thesis. I have a degree in this and I'm confused.
February 19, 2025 at 4:22 AM
I'm hung up on the presentation of this one because I don't think I'm gonna do for substructural typesystems what Rust did for affine typesystems; this is a six-pager and that's not the point. But I really need to do better at presenting it than this thesis. I have a degree in this and I'm confused.
Many problems dependent types solve can be tackled pragmatically with alternatives. Rust's const generics or C++'s non-type template parameters offer strong compile-time guarantees without full dependent types. #TypeSystems 3/6
November 3, 2025 at 5:00 PM
Many problems dependent types solve can be tackled pragmatically with alternatives. Rust's const generics or C++'s non-type template parameters offer strong compile-time guarantees without full dependent types. #TypeSystems 3/6
Elixir's gradual type system adoption is a hot topic. It allows developers to integrate types at their own pace, balancing immediate productivity with long-term robustness. A flexible path, but some debate its impact on broader adoption. #TypeSystems 2/6
October 18, 2025 at 10:00 PM
Elixir's gradual type system adoption is a hot topic. It allows developers to integrate types at their own pace, balancing immediate productivity with long-term robustness. A flexible path, but some debate its impact on broader adoption. #TypeSystems 2/6
New blog post: my understanding of GADTs (Generalized Algebraic Data Types) in OCaml.
durwasa-chakraborty.github.io/posts/gadts-...
#OCaml #TypeSystems #FunctionalProgramming #ProgrammingLanguages
durwasa-chakraborty.github.io/posts/gadts-...
#OCaml #TypeSystems #FunctionalProgramming #ProgrammingLanguages
GADTs in OCaml: The Type-Level Superpower Your Other Languages Wish They Had
As someone with a standing interest in languages, I find myself perpetually on the lookout for the features that belong to one language alone. Every language has its flagship concepts - words that res...
durwasa-chakraborty.github.io
June 30, 2026 at 1:39 AM
New blog post: my understanding of GADTs (Generalized Algebraic Data Types) in OCaml.
durwasa-chakraborty.github.io/posts/gadts-...
#OCaml #TypeSystems #FunctionalProgramming #ProgrammingLanguages
durwasa-chakraborty.github.io/posts/gadts-...
#OCaml #TypeSystems #FunctionalProgramming #ProgrammingLanguages
Does this mean that sound typesystems are cool again? http://news.dartlang.org/2017/01/sound-dart-and-strong-mode.html
Sound Dart and strong mode
As of the 1.19 release, Dart supports an optional mode, c...
news.dartlang.org
December 1, 2024 at 7:59 AM
Does this mean that sound typesystems are cool again? http://news.dartlang.org/2017/01/sound-dart-and-strong-mode.html
I wish my mouth was big enough to vocalize "fuck a typesystem" while also saying "bro have you seen how sophisticated Scala's type system is? OMGGG" while believing in my heart of hearts: "The C99 typesystem really leaves a lot to be desired and it being the best one is a strike against typesystems
June 17, 2025 at 10:44 PM
I wish my mouth was big enough to vocalize "fuck a typesystem" while also saying "bro have you seen how sophisticated Scala's type system is? OMGGG" while believing in my heart of hearts: "The C99 typesystem really leaves a lot to be desired and it being the best one is a strike against typesystems
I don't even like typesystems bcuz they distort and muddy the waters between who is the top and who's the bottom and its always between the programmer and whatever part of Scala it is that makes SBT good while somehow being worse than maven which sounds anatomically impossible bcuz it is #lein
June 17, 2025 at 10:29 PM
I don't even like typesystems bcuz they distort and muddy the waters between who is the top and who's the bottom and its always between the programmer and whatever part of Scala it is that makes SBT good while somehow being worse than maven which sounds anatomically impossible bcuz it is #lein
It really brings me joy that I'm in a place where I'm not a clojure dev endlessly talking about typesystems being shit from some kind of "lisp is better than all this bullshit" ivory tower but rather frok writing interface files making all the types we need for C, Objective-C, and Modula-2
June 10, 2025 at 1:50 AM
It really brings me joy that I'm in a place where I'm not a clojure dev endlessly talking about typesystems being shit from some kind of "lisp is better than all this bullshit" ivory tower but rather frok writing interface files making all the types we need for C, Objective-C, and Modula-2