Leave these fields empty (spam trap):
You can leave this blank to post anonymously, or you can create a Tripcode by using the format Name#Password
[i]Italic Text[/i]
[b]Bold Text[/b]
[spoiler]Spoiler Text[/spoiler]
>Highlight/Quote Text
[pre]Preformatted & Monospace Text[/pre]
[super]Superset Text[/super]
[sub]Subset Text[/sub]
1. Numbered lists become ordered lists
* Bulleted lists become unordered lists


its not about the language, just the logic

- Mon, 24 Oct 2016 08:33:25 EST CRO9Njb3 No.36253
File: 1477312405053.jpg -(30129B / 29.42KB, 284x177) Thumbnail displayed, click image for full size. its not about the language, just the logic
Fuck moving shit from one language to another.

building a universal translator for this shit.
so im going to pass everything to, lets say a .txt file since it's universal as shit.

what's a comparable attempt at this? I need to see where they stopped on the path of it being a truely universal translation.
Lydia Pevingridge - Mon, 24 Oct 2016 08:35:51 EST CRO9Njb3 No.36254 Reply
nb issue is longer than that,

the syntax is a matter of what line it would be stored as a raw value, array, dict. even whole Boolean statements could be covered. im just missing some aspect to the logic.
Lillian Clunderhidging - Mon, 24 Oct 2016 11:50:25 EST 9QSfnS0r No.36255 Reply
Deep learning / computational neural networks most likely.

Your most likely issue is: You have plenty of code to train the network to any one language but deriving equivalent operations not so much.

I'd start with the code from
and see what that gives you. Since each of the programs there has to not only arrive at the same output they also have to use the same algorithm.

You could use that and a natural language description as a starting point to be later be able to generate more cross language training data based on a bunch of github projects.
Jarvis Sangerbanks - Mon, 24 Oct 2016 16:19:54 EST Wtf4I+m0 No.36258 Reply
I remember being 18.
Phoebe Sammerstone - Mon, 24 Oct 2016 17:10:28 EST xLz8+Ihk No.36259 Reply
I used to work for a modernization/migration company.
We would take your existing source code, parse it into a language agnostic model, then convert it to a language specific model, then generate into the new language. For example we'd parse COBOL, put it in a transitional model, then translate that model to any other language you wanted, i.e. C++. The process seems like voodoo, and it basically fits your idea of a universal language. Trans-piling is something different.
Hannah Denkinfeck - Tue, 25 Oct 2016 21:12:18 EST fI0+sbkd No.36260 Reply
You do realize that all programs ultimately run on one of three(-ish) instruction sets, right? And that all three of those could be mapped to a turing machine?

What problem are you trying to solve?
Basil Blollypadging - Fri, 28 Oct 2016 22:41:35 EST gA3E5l0l No.36268 Reply
I am inclined to say that this is impossible. First of all, some languages have referential transparency while some do not, so any universal translator will need to "understand" what it's translating in a relatively deep way. For anything that touches memory directly you have not one model of computation but many depending on what type of machine the program is being run on and what state it's in. For purely functional languages you could spew out an equivalent program in a different purely functional language, but that program might not have anywhere near the same meaning to a human programmer, since in general there are many sources that compute the same function.

Actually, this might just reduce to the halting problem. If I had a universal program translator then it would send source code for a program that does not halt to source code for another (equivalent) program that does not halt. However, we know by the halting problem that there is no finite-length program which can make this determination in a finite amount of time, so the universal translator cannot exist.

Report Post
Please be descriptive with report notes,
this helps staff resolve issues quicker.