[slimygrapes]

The criticism re: purity is silly. It can still be useful to design a language that is “mostly pure” because you’re mostly not trying to fuck your self over. For example OCaml is “mostly pure” and it’s tremendously beneficial in terms of writing correct software, regardless of the IO escape hatch or the ‘a ref explicit mutability etc.

(I don’t care for V one way or the other though, and most of the results in the post are clearly bad, I just find this particular point to be absurd).

[mawfig]

To be clear, there's nothing wrong with impure programming languages. I use them every single day. My complaint is that V claims to be pure and is actually impure. To my knowledge, OCaml makes no claim to functional purity.

[Tozen]

"ML-derived languages like OCaml are "mostly pure". They allow side-effects through things like references and arrays, but by and large most of the code you'll write will be pure functional because they encourage this thinking. Haskell, another functional language, is pure functional. OCaml is therefore more practical because writing impure functions is sometimes useful."

https://ocaml.org/docs/functional-programming

> My complaint is that V claims to be pure...

"V is not a purely functional language however." (from V documentation)

https://github.com/vlang/v/blob/master/doc/docs.md#pure-func...

[mawfig]

"A pure function is one without any side-effects."

https://ocaml.org/docs/functional-programming

"V functions are pure by default, meaning that their return values are a function of their arguments only, and their evaluation has no side effects (besides I/O)."

https://github.com/vlang/v/blob/master/doc/docs.md#pure-func...

It sure sounds to me like V is trying to claim that functions can be pure while still performing I/O.

[creata]

Yeah, "no side effects (besides I/O)" is so absurd it's funny.

[slimygrapes]

Which means you are conceiving a strawman to mock.

The point is “no side effects other than IO” is a useful observation to make about how a language feels to use and the kinds of programs it encourages one to write.

Whether V is a good example of such a language I’ve no idea.

[amedvednikov]

Exactly, that was the idea. The wording was already changed a couple of weeks ago to more clear on the home page. Will update the docs as well.

[Tozen]

V's documentation is making it clear that it's not a purely functional language, so there should not be any such expectations. It also gives it's interpretation of what a pure function is, which other languages do as well, and there are various opinions about this.

You appear to be unbothered by OCaml describing itself as "mostly pure", whatever such "marketing" truly means, but seem to take strong offense to V. Seems like "mostly pure", in either case, should be equally acceptable or generate equal outrage.

[ptx]

This thread appears to be conflating two different things: whether a language is a "purely functional language" and whether the language's functions are "pure functions".

Presumably a purely functional language would, it seems to me, be exclusively (i.e. purely) functional, i.e. the functions in the language would always be pure. A language that has pure functions and impure functions, on the other hand, would not be "purely functional" as it contains impure parts, but it could still contain pure parts as well.

The issue with V seems to be that, although everyone agrees that it contains impure parts and thus is not a "purely functional language", the supposedly "pure functions" are also not actually pure, according to some reasonable definitions of purity.

[xigoi]

V's documentation says that V's functions are pure by default, but they aren't.