Comment by leafboi | Hacker News Reader

Comment by leafboi | Hacker News Reader

Parent Full thread

>You chose a definition among many and then dismissed the submission as "confused" because it does not meet your definition.

My definition? I quoted wikipedia. A lot of people use "my definition" Basically under "your" definition C++ and go might as well be a functional programming language.

>Considering that ML is one of the most impactful functional programming languages both in theory and practice, I think it's right to call that absurd and dogmatic. If you take it as a personal attack that I called your argument absurd, I don't know what to say. Maybe grow a little bit of skin?

Why grow skin? I literally didn't care. I just decided to call you junior because you were acting like one. It seems like you cared more. I was just explaining to the other guy why I called you a junior. I recognize the attack but I literally don't care, but that doesn't mean I won't respond or address it.

>I don't need to provide a definition to find a flaw in yours. I did not come into this submission looking to smugly dismiss the topic at hand with my obviously irrelevant preferences for terminology. You did.

My Definition doesn't have flaws. Like I said, if I expand it to encompass mutability (as you have) then you can place Go and C++ on the functional pedestal. I haven't used ML but clearly a language like ML only allows mutability as a small exception that's rarely used otherwise there's no point to classify ML as functional.

But let's get back on topic. You literally said that no PL theorist would agree with me. But look at wikipedia. Apparently many do. So you're wrong.

I don't take wikipedia to be authoritative about anything but I'll bite.

> Functional programming is sometimes treated as synonymous with purely functional programming, a subset of functional programming which treats all functions as deterministic mathematical functions, or pure functions.

Well, obviously in this context the author of the submission wasn't using that definition. So you coming in and dismissing the conversation because sometimes people mean "pure functional programming" when they say "functional programming" is clearly erroneous. I know when people make this conflation, often when they're talking about Haskell. When talking about ML, no one would take that synonym. Context matters.

I just looked, you provided a set theoretic definition of functions, the wikiwand one. You chose a poor definition for the topic at hand. In most mathematical settings it would be a perfectly fine definition, but not here. Context matters.

Here are the problems with it: It assumes an interpretation in sets when there are a plurality of interpretations of functions qua functional programming which are not compatible with set theoretic models. For example, functions in homotopy type theory have higher dimensional structure than just being elements of a set. Functions in domains have more structure than just the extensional mapping from inputs to outputs. Your chosen definition also includes non-computable functions which are simply not admissible in this context.

So your definition is simultaneously too restrictive and too lax.

You seem to be confused by the existence of multiple definitions and how they are appropriate in context.

>I don't take wikipedia to be authoritative about anything but I'll bite.

No but it's more authoritative than a random person on the internet (you). For things like this it's good enough. I'm not writing a paper here.

>Well, obviously in this context the author of the submission wasn't using that definition.

>Context matters.

>In most mathematical settings it would be a perfectly fine definition, but not here.

In this context it's also perfectly fine. OP did not make a distinction between subroutine or function. He also uses the term higher order function wrong. It shows two things. He does not know what a higher order function is and he does not know what a function is.

>Here are the problems with it: It assumes an interpretation in sets when there are a plurality of interpretations of functions qua functional programming which are not compatible with set theoretic models. For example, functions in homotopy type theory have higher dimensional structure than just being elements of a set. Functions in domains have more structure than just the extensional mapping from inputs to outputs. Your chosen definition also includes non-computable functions which are simply not admissible in this context.

This is the garbage pedantry that I warned you about earlier. I'm choosing the generally accepted academic definition. Hopefully you're not one of those people who can't comprehend "generally accepted" without an expose into axioms and different obscure branches of math.

>You seem to be confused by the existence of multiple definitions and how they are appropriate in context.

Find the generally accepted definition in wikipedia that this person is using (Hint: no actual technical definition exists). Obviously he's using "function" in place of procedure or subroutine. But then he uses the term "Higher Order" to imply an actual functions instead of procedures.

Not only does he use "Higher Order" incorrectly but his usage of the term indicates he doesn't know what a function is and the difference between a function and a procedure.

Yeah I'm not going to reply to your shit any more. HN really needs to change the policy of not letting parent posters vote others down.

> This is the garbage pedantry that I warned you about earlier. I'm choosing the generally accepted academic definition. Hopefully you're not one of those people who can't comprehend "generally accepted" without an expose into axioms and different obscure branches of math.

I think you should look into domain theory it is one of the most celebrated discoveries pertaining to the semantics of functional programming languages. And maybe you'll understand that from a programming point of view, what a function is is much more flexible and subtle than you think.

> Yeah I'm not going to reply to your shit any more. HN really needs to change the policy of not letting parent posters vote others down.

Would it help if I pretended that you downvoted me? Every time I see my imaginary internet points I will subtract 5.

No you should just stop voting me down. I'm literally going to go negative soon, I just joined. I literally cannot talk to you if you continue voting me down. It's bad form.

>I think you should look into domain theory it is one of the most celebrated discoveries pertaining to the semantics of functional programming languages. And maybe you'll understand that from a programming point of view, what a function is is much more flexible and subtle than you think.

Sure I could. But "context matters." This is not the topic of the conversation. There's a level of depth we're operating at here. Very few people understand "domain theory" and you shouldn't assume as much when you communicate with anyone.

Admittedly I downvoted this one: https://news.ycombinator.com/item?id=24438137

Yet again though:

"Random person (you)" "he does not know what a function is." "Yeah I'm not going to reply to your shit any more."

...give a good reason

---

Anyways, leaving this thread. I certainly learned something from it, I hope you did too! : )

I now noticed you're at -1, went ahead and gave you an upvote. Hopefully you can still post, and be a bit more aware.

--

Also, great work trying to explain bitdizzly. Def enjoyed reading your explanations, and will explore some the references you popped in

There's nothing for me to learn here other than domain theory which is very advanced math and how people behave.

If you follow the conversation you'll see that the entire argument turned into a battle of definitions with me turning to a very generally accepted definition of a function and bitdizzy turning to using more obscure mathematical definitions of "function" and arguments about semantics what context is relevant to what definition of "function." Whatever definition you choose to believe domain theory looks like very advanced stuff and clearly outside the scope of your average technical conversation.

Either way, I didn't have time to respond to your other comment below. But I will now.

Don't worry about the upvote thing. It's just for this thread I can switch accounts.

I'm not downvoting you. I can't downvote anyone. I don't think I have enough points to downvote. If I didn't want to talk to you I wouldn't be replying!

Reply on news.ycombinator.com