Hacker News Reader

Comment by logical_proof | Hacker News Reader

I think you just made an equivocation as those Time.now’s are not the same

  A = Time.now
  A != Time.now
  A == A

Just because Time.now looks like the same symbol it is not as it is an impure function that never returns the same result.

View on HN

We are not talking about the same computation. What you think I mean is [1]. What I actually mean is [2].

  [1] pry(main)> eval("a != a")
  NameError: undefined local variable or method `a' for main:Object from (pry):1:in `eval'
  [2] pry(main)> eval("Time.now != Time.now")
  => true

I am not assigning the function to a variable - I am lazy-evaluating it.

It's not equivocation because [2] is expressed exactly the same way in English. Now is not now. It is trivially true.

When I use a word, Humpty Dumpty said in rather a scornful tone, it means just what I choose it to mean — neither more nor less

I disagree, because in both programming languages and in natural language, "now" is not a constant: it's a function that returns a different value depending on when it's called.

While the syntax is "now", what you're really saying is now(Ti) for some i.

So the sentence, in both natural language and most programming languages becomes:

    now(T0) != now(T1)

Which is easier to understand, with all arguments made explicit.

Now, some "clever" programming language might automatically translate this from

    now != now

    a := now(T0)
    a != a

Which will be surprising to the programmer in most cases (but might be what's desired in others).

The important thing is that "now" is not a comstant, but a function.

Yes! "Now" is not a constant. It's a function. It's a function that takes no arguments. It's also a function that can be lazy-evaluated.

  f() != f()
  A != A

You seem to be agreeing with me but your response to me starts with the English phrase "I disagree".

I disagree, that's the thing: "now" is a function that takes one implicit argument: time.

So "now" is a trickery to hide the fact you're really saying "now(Ti)", for varying i.

I'm not tripping over syntax, but I think you are:

You think you're saying

    A != A

When you're really saying

    A1 != A2

(Or rather

    now(t0) != now(t1)

)

And of course the second form is unsurprising.

You are talking about a different function.

The function I am talking about doesn't take any arguments.

    [1] pry(main)> Time.now(1)
    ArgumentError: wrong number of arguments (given 1, expected 0)

> The function I am talking about doesn't take any arguments.

All impure functions have an implicit dependency on (some subset of) the state of the universe when they execute; in mathematics and logic something with this kind of dependency is typically noted by parameterizing it by time when the actual precise nature of the dependency is outside of the scope of the immediate analysis (because a dependency on the state of the universe at the time of execution is equivalent to a dependency on the time of execution with a hidden function mapping time to state of the universe.)

That impure functions methods are in many programming languages syntactically not distinguished from pure ones does not change the logic here.

That Time.now(at time t0) is distinct from Time.now(at time t1) is... unsurprising, and not an example of A != A in logical terms.

By the way... your conception of a pure function is incoherent.

Parameters ARE inputs. Results ARE outputs. Renaming them to something else is just obscuring the fact that you are still talking about a function with side-effects.

So using Python as my model of computation this is a pure function, but it does nothing:

   def f(): pass

This is NOT a pure function. It has outputs.

   def g(): return True

This is NOT a pure function. It mandates input.

   def h(x): pass

This is NOT a pure function - it mandates input and produces output:

   def j(x): return x

This is a pure function. It calculates the answer to 2+2, but doesn't tell you what it is.

   def k(): 2+2

If pure functions take no input and produce no output, what is there left to analyse? What is Logic and Mathematics talking ABOUT?

Well, it seems our disagreeme is obvious.

You only want to talk about pure functions. I want to talk about ALL functions.

So lets define the function "pure" such that

  pure(f) -> True when f is a pure function
  pure(g) -> False when g is an impure function

What would you say is the truth-value of pure(pure) ?

https://en.wikipedia.org/wiki/Impredicativity

dragonwriter wrote:

> "All impure functions have an implicit dependency on (some subset of) the state of the universe when they execute"

You wrote:

> "You only want to talk about pure functions. I want to talk about ALL functions."

Find the error.

You know what... let me take your red herring away from you. Forget the Time.now function. Here's a !=a as a "pure" function.

   In [1]: class A:
   ...:     def __eq__(self, other):
   ...:         return False
   ...:
   In [2]: a = A()
   In [3]: assert a != a
   In [4]: assert a == a
   AssertionError

Will you now permit me to speak freely about A != A without you moralising about my reasoning?

Will you now recognise my right to free speech and free thought, or must I fall in line?

>Find the error.

Still waiting for you to implement equal() as a "pure" function.

We can discuss what you mean by "error" once we have a referent (that's not just in your head) to talk about.

You know what the state of the universe is when you do logic?

Tell us more! I didn't believe in omniscience until now.

Time.now(t1) and Time.now(t2) is meaningless in this context because t1 and t2 are unbound variables. What are you passing to your function? Where is it coming from?

My function says you are overlooking something.

  [1] pry(main)> Time.now(t1)
  NameError: undefined local variable or method `t1' for main:Object from (pry):1:in `__pry__'
  [2] pry(main)> Time.now(1)
  ArgumentError: wrong number of arguments (given 1, expected 0) from (pry):2:in `now'

If you're familiar with the literature of programming languages, you'll see references to the "world" or the "universe" all the time. That's the kind of universe we mean in this discussion, not the one Carl Sagan talked about.

As for the rest: please provide the implementation of your Time.now function and I'll tell you its implicit parameters. If you're not talking about an actual function that can be run by a computer, then please tell me so and I'll drop it -- I'm not interested in hypothetical functions tautologically defined not to be equal to themselves.

>please provide the implementation of your Time.now function and I'll tell you its implicit parameters.

You are missing the point about extensional and intensional properties of functions.

If you want to look "inside" my Time.now function then maybe you should also look "inside" your equality function and find its implicit parameters?

Here's some Python boiler plate - implement equality and make the assertions pass.

  def f():
    return 1

  def equal(x,y): 
    # Implement equality

  assert not equal(1,2)
  assert equal(1,1)
  assert equal(f, f)

>If you're not talking about an actual function that can be run by a computer

As far as I can tell you are not talking about such a function either when you speak about equality, but I'll reserve judgment until you produce an implementation.

>I'm not interested in hypothetical functions tautologically defined not to be equal to themselves.

OK, but you are interested in hypothetical functions tautologically defined to be equal to themselves.

That's perfectly fine - we are interested in different tautologies.

You want to say equal(f,f) is true.

I want to say not equal(f,f) is true.

We want to say different things.

It doesn't take an explicit argument, but time is definitely an implicit argument of your function; how else do you think the function is implemented?

We are talking about function declaration not implementation, right?

The expression A = A can be trivially re-written in another notation as equal(A,A). The reason you don't question the implementation of the equal() function is because it's declared as being true however it's implemented.

So, I declare f() != f() as being true. I also declare that f() takes no arguments. All I am doing is translating (transpiling?) English into Ruby.

"Now is not now" => "Time.now != Time.now"

Talking about function arguments is not even in point. What's the argument to now() in English?

> All I am doing is translating (transpiling?) English into Ruby.

Incorrectly.

> "Now is not now" => "Time.now != Time.now"

Sure, “Now (pointedly looks at watch) is not now (pointedly looks at watch)” is true, and the equivalent of the Ruby. Ruby expressions lack the implicit simultaneity that English sentences wthout some contextual signal of explicit time dependency have, and your mistranslation ignores this.

Where is the "explicit simultinaity" in looking at your watch twice?

The English and Ruby terms map 1:1.

Now(Time.now) is not(!=) Now (Time.now)

Why do you call the correct "incorrect"?

> Where is the "explicit simultinaity" in looking at your watch twice?

Nowhere; that's the contextual signal that defeats the implicit simultaneity the English sentence would have without that body language.

So you looked at your watch twice at the same time?

The English sentence "Now is not Now" describes my thought process (1st person view), not my body language (3rd person view).

My thought process occurs over time. Exactly like the Ruby process.

Reply on news.ycombinator.com