3 ms·

what "advanced" even mean? All those languages, they don't do something that others are not capable of(not mention that all the jvm ones just produce bytecode o

what "advanced" even mean? All those languages, they don't do something that others are not capable of(not mention that all the jvm ones just produce bytecode over the same vm)...different development principles is not a reason to call a language advanced.

MichaelBurge 10y ago
Haskell and ML at least have advanced type systems. Since types are erased during compilation, you can't reason about type systems in terms of their impact on the generated code. An example of a simple type system is C with its structs. Haskell also has lazy evaluation, which is very difficult to do in C. Obviously you could encode your own trampolines, but at that point you might as well be coding in a Turing tar pit language "where anything is possible but nothing of interest is practical". Pattern-matching and ADTs are also usually difficult to do in other languages. Most of the benefits of something like Haskell only show up on larger projects, so it's not meaningful to point to specific language features as a point against it. I can do just about anything in C. The type systems you probably can't do; you'd have to embed a language and that doesn't count.
superuser2 10y ago
* A lot of mainstream languages are syntactic skins on top of substantially the same semantics and skill in one is easily transferrable. A person skilled in C should be able to understand a Java program, but may find similarly high-quality, idiomatic Haskell totally inscrutable. * The number and complexity of concepts with which you should be familiar are much higher. To tread water in most imperative languages, you need to find if/else, while, for, assignment, function call, and function definition. To tread water in these languages, you will generally need more. * The idioms provided by these languages provide for and encourage code that is more expressive (says more with fewer characters) but may be more difficult to think about and write. * All languages are equivalent in the sense that you can simulate them all by each other, i.e. by writing a compiler from X programs to lambda calculus expressions, and a lambda calculus interpreter in Y. This result does not imply that you can always express the same concepts in a reasonable, space-efficient manner. I do have some nits to pick with the list, though. I would consider C++ advanced, in that there is a lot you need to know and think about to work with it. Scheme itself is very simple: the entirety of its syntax and semantics can be learned by a laymen in an hour, and by the end of a semester first-year computer science students can write their own Scheme interpreters in Scheme. Its complexity is an emergent property of the kinds of abstractions it supports, not inherent to the size and complexity of its design like Haskell or Scala.

MichaelBurge 10y ago

Haskell and ML at least have advanced type systems. Since types are erased during compilation, you can't reason about type systems in terms of their impact on the generated code. An example of a simple type system is C with its structs. Haskell also has lazy evaluation, which is very difficult to do in C. Obviously you could encode your own trampolines, but at that point you might as well be coding in a Turing tar pit language "where anything is possible but nothing of interest is practical". Pattern-matching and ADTs are also usually difficult to do in other languages. Most of the benefits of something like Haskell only show up on larger projects, so it's not meaningful to point to specific language features as a point against it. I can do just about anything in C. The type systems you probably can't do; you'd have to embed a language and that doesn't count.

superuser2 10y ago

* A lot of mainstream languages are syntactic skins on top of substantially the same semantics and skill in one is easily transferrable. A person skilled in C should be able to understand a Java program, but may find similarly high-quality, idiomatic Haskell totally inscrutable. * The number and complexity of concepts with which you should be familiar are much higher. To tread water in most imperative languages, you need to find if/else, while, for, assignment, function call, and function definition. To tread water in these languages, you will generally need more. * The idioms provided by these languages provide for and encourage code that is more expressive (says more with fewer characters) but may be more difficult to think about and write. * All languages are equivalent in the sense that you can simulate them all by each other, i.e. by writing a compiler from X programs to lambda calculus expressions, and a lambda calculus interpreter in Y. This result does not imply that you can always express the same concepts in a reasonable, space-efficient manner. I do have some nits to pick with the list, though. I would consider C++ advanced, in that there is a lot you need to know and think about to work with it. Scheme itself is very simple: the entirety of its syntax and semantics can be learned by a laymen in an hour, and by the end of a semester first-year computer science students can write their own Scheme interpreters in Scheme. Its complexity is an emergent property of the kinds of abstractions it supports, not inherent to the size and complexity of its design like Haskell or Scala.