6 ms·

One of the issue with structural types is compiler error messages. By example, C++ templates are structural types and error messages are really hard to decipher

One of the issue with structural types is compiler error messages. By example, C++ templates are structural types and error messages are really hard to decipher. Recent versions of Java (Java 8+) avoid structural types but provide structural conversions + inference. We will see with the recent introduction of named records/tuples goes in the same direction or not.

ncmncm 5y ago
C++ templates used to be structural types. Since C++20, C++ supports meta-type constraints on templates. Of course, backward compatibility means those are optional, but the designer of the template decides. Existing libraries are all structural, but new ones will rely more on nominal meta-typing, and error messages for those are clean. There will be pressure to retrofit older libraries, so that error messages from misuses of those get nicer too. Gripes from users who had successfully abused those libraries will be heard.
- pjmlp 5y ago
  Partially, because they only work for the caller, on the implementation itself, one can still make use of structural typing on top of concept constrained parameters and the compiler will be silent on that until the template gets instatiated and the traditional compiler vomit ensues.
  - ncmncm 5y ago
    You are just talking about when they are applied, not what can be done. Nothing was removed from C++: you can still write templates with structural constraints, or trait constraints, or any mix. In Rust, a generic is only right or wrong, in isolation. In C++, how it behaves, including whether it compiles, depends on how you choose to use it. That makes C++ templates strictly more powerful than Rust facilities. Rust could relax some restrictions, selectively, to gain that power, and might someday. Preventing the "compiler vomit" is a main selling point of Concepts, and it works. It is still possible to subvert it, if you work at that (then you are no worse off than before). So, don't.
    - pjmlp 5y ago
      I am talking about what we got is concepts light, and not the full deal. When one works with teams our options don't work in isolation. As for concepts working in preventing compiler vomit, there is still plenty of work to do in 2021.
      - ncmncm 5y ago
        The "full deal" turned out to be undesirable, and undesired. There is always more work to do. That is not a failing, it's just life.
        pjmlp 5y ago
        The full deal failed due to the usual ISO politics, in the end it was concepts light or nothing. That is what defines modern C++, a standard full of political compromises, slowly becoming a niche language.
        ncmncm 5y ago
        It failed because it turned out not to be usable, in practice. Getting something equally ambitious and also usable was a research project. Blaming ISO politics is a way to avoid need to understand unpleasant technical details. You wish C++ were becoming a niche language, but in fact its usage is still growing by leaps and bounds, as it has been continually since C++11 came out. Every week more people pick up C++ than the whole population now employed coding that other language. That will be true next year, too.
        pjmlp 5y ago
        You wish C++ were as widely used across the whole OS stack as back in its glory 1990's days. The fact is that it has become a niche language for GPGPU programing, OS drivers and embedded standards like AUTOSAR, everywhere else another language takes the crown jewels, iOS, Android, macOS, ChromeOS, Windows Apps, Web, cloud computing infrastructure,... where is your growing C++? So much that Apple and Google are now rather focusing on their own languages, with C++17 being good enough for their own purposes, hence why clang is trailing its ISO C++20 compliance. When everyone brings their own agenda to the table regarding implementation details it is politics, not only concepts, contracts, ABI breaks (god forbid!), reflection, networking, graphics,...., definitly politics. But hey, it is growing on GPGPU, HPC, Machine Learning libraries for Python, LLVM/GCC implementation language, so there is that.
- Kranar 5y ago
  Before C++20 templates were untyped. Since C++20 it's possible to define structural types for templates via concepts. Concepts are an explicitly structural type system for templates. Nominal typing of generic code would be what most other languages that supports generic programming use. For example Rust has traits where one defines a named trait and must explicitly specify that a type satisfies a trait by name. In C++ there is no such means, a concept does strictly structural checks (via constraints) to determine if a template satisfies a concept, no names are involved. In other words, a type T satisfies a concept C defined by a constraint R if and only if a T can be substituted for C's placeholder type during the evaluation of R.
  - gpderetta 5y ago
    Everything you said is very much true, but there is a kind of nominal typing even in template land via (partial) specialization of special trait classes. For an example see std::tuple_size/tuple_element and their relation with (ironically) structured binding.
    - ncmncm 5y ago
      Right: C++ offers a continuum between pure structural typing and nominal typing. Author chooses according to need. Failing to support structural typing is not an advantage.
      - gpderetta 5y ago
        Yes, as usual, name a language feature and it is probably somewhere in C++ :) (mind, I'm not complaining).
tsimionescu 5y ago
> By example, C++ templates are structural types and error messages are really hard to decipher. To be fair, C++ templates are not structural types, they are (together with C++ function overload resolution) an (almost) purely functional untyped programming language, producing C++ programs as output.

ncmncm 5y ago

C++ templates used to be structural types. Since C++20, C++ supports meta-type constraints on templates. Of course, backward compatibility means those are optional, but the designer of the template decides. Existing libraries are all structural, but new ones will rely more on nominal meta-typing, and error messages for those are clean. There will be pressure to retrofit older libraries, so that error messages from misuses of those get nicer too. Gripes from users who had successfully abused those libraries will be heard.

pjmlp 5y ago
Partially, because they only work for the caller, on the implementation itself, one can still make use of structural typing on top of concept constrained parameters and the compiler will be silent on that until the template gets instatiated and the traditional compiler vomit ensues.
- ncmncm 5y ago
  You are just talking about when they are applied, not what can be done. Nothing was removed from C++: you can still write templates with structural constraints, or trait constraints, or any mix. In Rust, a generic is only right or wrong, in isolation. In C++, how it behaves, including whether it compiles, depends on how you choose to use it. That makes C++ templates strictly more powerful than Rust facilities. Rust could relax some restrictions, selectively, to gain that power, and might someday. Preventing the "compiler vomit" is a main selling point of Concepts, and it works. It is still possible to subvert it, if you work at that (then you are no worse off than before). So, don't.
  - pjmlp 5y ago
    I am talking about what we got is concepts light, and not the full deal. When one works with teams our options don't work in isolation. As for concepts working in preventing compiler vomit, there is still plenty of work to do in 2021.
    - ncmncm 5y ago
      The "full deal" turned out to be undesirable, and undesired. There is always more work to do. That is not a failing, it's just life.
      - pjmlp 5y ago
        The full deal failed due to the usual ISO politics, in the end it was concepts light or nothing. That is what defines modern C++, a standard full of political compromises, slowly becoming a niche language.
        ncmncm 5y ago
        It failed because it turned out not to be usable, in practice. Getting something equally ambitious and also usable was a research project. Blaming ISO politics is a way to avoid need to understand unpleasant technical details. You wish C++ were becoming a niche language, but in fact its usage is still growing by leaps and bounds, as it has been continually since C++11 came out. Every week more people pick up C++ than the whole population now employed coding that other language. That will be true next year, too.
        pjmlp 5y ago
        You wish C++ were as widely used across the whole OS stack as back in its glory 1990's days. The fact is that it has become a niche language for GPGPU programing, OS drivers and embedded standards like AUTOSAR, everywhere else another language takes the crown jewels, iOS, Android, macOS, ChromeOS, Windows Apps, Web, cloud computing infrastructure,... where is your growing C++? So much that Apple and Google are now rather focusing on their own languages, with C++17 being good enough for their own purposes, hence why clang is trailing its ISO C++20 compliance. When everyone brings their own agenda to the table regarding implementation details it is politics, not only concepts, contracts, ABI breaks (god forbid!), reflection, networking, graphics,...., definitly politics. But hey, it is growing on GPGPU, HPC, Machine Learning libraries for Python, LLVM/GCC implementation language, so there is that.
Kranar 5y ago
Before C++20 templates were untyped. Since C++20 it's possible to define structural types for templates via concepts. Concepts are an explicitly structural type system for templates. Nominal typing of generic code would be what most other languages that supports generic programming use. For example Rust has traits where one defines a named trait and must explicitly specify that a type satisfies a trait by name. In C++ there is no such means, a concept does strictly structural checks (via constraints) to determine if a template satisfies a concept, no names are involved. In other words, a type T satisfies a concept C defined by a constraint R if and only if a T can be substituted for C's placeholder type during the evaluation of R.
- gpderetta 5y ago
  Everything you said is very much true, but there is a kind of nominal typing even in template land via (partial) specialization of special trait classes. For an example see std::tuple_size/tuple_element and their relation with (ironically) structured binding.
  - ncmncm 5y ago
    Right: C++ offers a continuum between pure structural typing and nominal typing. Author chooses according to need. Failing to support structural typing is not an advantage.
    - gpderetta 5y ago
      Yes, as usual, name a language feature and it is probably somewhere in C++ :) (mind, I'm not complaining).

tsimionescu 5y ago

> By example, C++ templates are structural types and error messages are really hard to decipher. To be fair, C++ templates are not structural types, they are (together with C++ function overload resolution) an (almost) purely functional untyped programming language, producing C++ programs as output.