Hacker Newsnew | past | comments | ask | show | jobs | submit | raoof's commentslogin

FRACTRAN has a goto instruction so it is not a total model of computation but my claim is that Ar is a total model of computation. maybe unrolled FRACTRAN is a total model of computation too, I have to think about that


if you don't want to melt your brain :D you just precompute the value of any functions f(x) for x<n otherwise you need to be very clever I explained my idea in the link


Don't need a language then, just compile down to "return 4". No instructions needed. Without user input it's not really useful. How about returning maximum value of user-privided array of int64 values? If array is static, just return final value and that's all. Smart compiler for the win!

P.S. precomputing even a single int64 space is going to be really really hard, 2^64 possible values is a lot


the only approach to practical programming with Ar is to have external dependencies on hard to compute functions until I implement them :)


That's why I said an arbitrary number, you can't just precompute the primality of every single number... And I'm afraid I don't really understand your proof of being able to do loops without any kind of looping construct in the language itself


(it seems like I can't reply to your last comment so I do it here) it depend how you want to display the result to the user, you can just compute a number that represent an html page showing the numbers from 1 to n


the language is obviously (at least to me) turing-complete in the limit of the program length. even if you don't buy my argument that it is also turing-complete without the limit notice that in ordinary programming you never need a function that is valid on all input because actual computers have time and memory limitation


I must just be missing something then. How would you write a program that prints out the numbers from 1 to n, for reasonable n? (representable by a computer)


I think the issue here is that the language doesn't have any form of user input, so all possible programs are fully static and thus can be represented without any loops or conditionals.

I don't agree that the language is turning complete due to the lack of any runtime dynamic aspect of the language.

It's akin to compiling a simple c program with loops unrolled and fully static variables and claiming the subset of generated assembly is turning complete.


Agreed.

> It's akin to compiling a simple c program with loops unrolled and fully static variables

Yes, and optimizing compilers take this to an extreme. It turns out that "computing" a factorial only requires my runtime language to have `mov` and `ret` instructions: https://godbolt.org/z/3Y6bGcsba


in my model it's the platform job to handle the user input and your program output you just write your function and the platform calls your program for example f(time,input_data,key_a,key_b,key_c,...) . if you are very very patient and don't care about the program length you can just compute every output of your program for every possible input. according to [this](https://web.archive.org/web/20230505022436/http://page.mi.fu...) paper a single loop and four basic arithmetic operation is enough to simulate a turing machine so my idea was that to unroll the loop and my conjecture is that we don't need the loop and only four operations is enough to compute all total computable functions


(now I can reply here, let me repeat it again) it depend how you want to display the result to the user, you can just compute a number that represent an html page showing the numbers from 1 to n


The question isn't as to how it will be represented, that's the boring part.

When given a number n, and you want to display them, how will you actually compute the numbers from 1 to n? If you precalculate a table - then how will you do n+1? What about n+2?


it takes a lot of time and patient to come up with a function that works for all n and also have a short length, let me try if I can solve this problem it may takes a few hours or days or maybe months I'll post the solution here so stay tuned. but if you want to make your life easier you can generate the function in a meta language so you can always be sure that your software does not have a halting problem


consciousness is like a joke, if you think it needs an explanation you already miss it. I struggled with myself to convince Dennett that his conscious so much that I ended up losing it myself I hope you don't make the same mistake that I did


consciousness is equivalent to the halting function h(p,n) where p is a physical process and n is time, we know how to turn a physical process into a virtual process but we don't know how to turn consciousness itself into a virtual process


if Reality!=0.111... :

  the Universe is computable ( the list of infinite binary strings between [0,1) is countable )
else:

  the Universe is non-computable ( the list of infinite binary strings between [0,1] is uncountable )
how do you solve this dilemma?


Isn't the list of infinite binary strings between [0,1) also uncountable?


if you put 1 out then there is no reason to believe it is uncountable because you can sort the list such that the diagonal is all 0s (if you know an argument that shows [0,1) is uncountable I'd be happy to know)


It is a superset of (0,1), which is uncountable because it is an open interval. All open intervals except the empty one are uncountable. What kind of proof would you like?

There is also no such thing as an uncountable list. Lists can be indexed by integers, which makes them all countable.


the proofs that I've read shows that the list of infinite binary strings between [0,1] are uncountable because the invert of the diagonal of the list is not on the list, do you know any proof that does not refer to the invert of the diagonal?


Well, if you remove one element from an uncountable set, then...


you can't prove something just by assuming it that would be trivial, if we can build or convince ourselves how to build a wormhole or a time machine then we can believe in Real numbers


What Koshkin says is correct and has already been proven by others. You want to get a book that covers sets, maps and cardinal numbers, and read it carefully if you are interested in this sort of stuff.


I can see that you can symbolically talk about the power set of (0,1) and say it's cardinality is bigger than the set itself (although I haven't studied sets and cardinalities deeply enough) but I can't see what Real numbers offers against the computable Universe hypothesis


thank you for the recommendation


that would get way too much time to reach an acceptable state for one person and creating your own game engine is much better


interesting in any sense


thanks for the idea but not that interesting beside I don't have access to RPi


didn't read the whole article but I have to say that having a non-mechanical process in the ontology of the Universe does not mean nonsense, it's not obvious that AI is not possible but also it's not obvious that we are just a body so why should we believe in a meaningless and boring universe rather than a meaningful and wonderful one and also if it turns out that there is no room for a non-mechanical process in the Universe then we can believe that consciousness is equivalent to existence meaning if the Universe seems meaningful and wonderful then it is really meaningful and wonderful no matter how much you try to explain the meaning and wonder away.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: