Paradoxes

[ChrisWill]

I'm a programmer but since playing ksp I've been researching physics a lot more. I enjoy math and physics especially theoretical stuff and paradoxes and started wondering if there are any programming paradoxes. If not could a programming paradox even exists considering we can just tailor things to are needs were in space and time we can't. I've done some research and come up empty the only thing close to a paradox can't really be considered a paradox in a traditional sense. Which is "The Singularity" in where computers and machines can upgrade and build newer and better versions of themselves and would it eventually get to perfection where it can't upgrade because there's nothing beyond perfection thus creating a perfect computer.

[Godot]

How would you measure "perfection" in a computer?

Considering the number of possible tasks and their different requirements, as well as the huge number of other possible attributes of a computer (like power consumption, heat signature, price, size, vulnerability towards radiation, MTbF of its components)

I don´t think that there can be a "perfect" computer. There can only be approaches that make the computer "perfect" one task (or "perfecting" one attribute), while making the other attributes of the computer (or its usability for other types of tasks) as good s possible under this constraint (but far from perfect).

Therefore I think a "perfect" computer would be a Paradoxon by itself

Edited October 19, 2013 by Godot

[DaveofDefeat]

Calculate pi.

[ChrisWill]

How would you measure "perfection" in a computer?

Considering the number of possible tasks and their different requirements, as well as the huge number of other possible attributes of a computer (like power consumption, heat signature, price, size, vulnerability towards radiation, MTbF of its components)

I don´t think that there can be a "perfect" computer. There can only be approaches that make the computer "perfect" one task (or "perfecting" one attribute), while making the other attributes of the computer (or its usability for other types of tasks) as good s possible under this constraint (but far from perfect).

Therefore I think a "perfect" computer would be a Paradoxon by itself

That's very true.

Calculate pi

Haha yeah reminds of this though http://www.numberworld.org/misc_runs/pi-10t/details.html

[Crush]

Paradoxes don't exist in reality.

Regardless of scientific discipline, you only encounter paradoxes in theory, and when you encounter them it proves that your theory is incorrect, incomplete or based on wrong information.

Take the liars paradox, for example ("Everything I say is a lie!"). The paradox proves that the data got to be wrong.

[K^2]

Any time you are dealing with logic programming, especially in knowledge bases, you may encounter what is strictly speaking a paradox, but these are just statements that reduce any knowledge base to a contradiction. If you are trying to build a knowledge base for some practical application, you should guard against these, because feeding a paradox to an inference-based KB clears the KB.

Of course, the way you check is also very simple. Any paradox reduces to a contradiction on its own, so all you have to do is reduce the statement being fed, and if it evaluates to false, you don't append it to a KB.

[Scotius]

In some works of classic sci-fi authors (like A. Clarke) exist 'paradox-commands' - used by protagonists to deal with annoying/obnoxious robots. Every part of such command is perfectly clear, but together they form a internally counteracting sentence, that will lock simple automaton in logic loop. For example "Move outside and let me see you" (translation might be a bit wrong, sorry) will cause a robot to start turning endlessly in one spot when he is trying to obey both commands at once: leave the room and still be seen by command-giver (which is physically impossible). Of course more advanced, self-aware androids are unaffected, just like most humans would be. I was always wondering if such thing as paradox-command would be possible? It seems a bit dumb to build a fairly complicated robot capable of following voiced orders, without installing safeties assuring machine will not be rendered useless by some random smart-mouth

[K^2]

I was always wondering if such thing as paradox-command would be possible?

Yeah, like I said above, a simple paradox locks up a simple logic system, but they are easy to filter for. So you just have to foresee that such instructions are possible, test for them, and have the contingency planned, like informing user that instruction cannot be performed.

[Crush]

In some works of classic sci-fi authors (like A. Clarke) exist 'paradox-commands' [...] For example "Move outside and let me see you" (translation might be a bit wrong, sorry) will cause a robot to start turning endlessly in one spot when he is trying to obey both commands at once

Wasn't that from a short story by Isaac Asimov?

[NGTOne]

In my view, no - there aren't any paradoxes that are strictly mathematical/logical. So far as I'm aware, all paradoxes arise from constructions of language, which by its nature is an imprecise and illogical thing (hence the paradox). Any purely logical processing entity (i.e. a computer or robot) would eventually reach a mathematical contradiction (this is a formal-logic term, with a precise meaning - think 2=3 to get the general idea). How it dealt with this contradiction would dictate its response (whether it crashes, freezes, or simply ignores it). Any halfway sane programmer would tell it to simply ignore the contradiction, as opposed to having it freeze or crash (it's like a divide-by-zero - a ridiculously basic condition that is dealt with trivially in ALL modern computer systems at a very low level).

[hugix]

One of the most popular paradoxes do not apply to computers. At least this is my interpretation.

->This sentence below this one tells the truth

->The sentence above this one lies

For us this is impossible because both meanings conflict. But a computer doesn't care about the meaning of the first sentence when it is computating the second. It just accepts the order to regard the first one as a lie. It does not care about it's content.

[Moach]

Windows 8 could be considered a paradox...

Really - Think about the fact that if only that... *cringe* "thing" had been released under the name of "Windows Touch" or something, and thus marketed as a new mobile platform, and not a supposed "replacement" for trusty old '7; M$ could have very well outdone Apple by lightyears, releasing the first-ever mobile OS that can run all and any existing windows software you could think of without any publisher-made adaptations....

well, that's not really how it happened now, was it? *where's the facepalm smiley?*

the line of thinking that led a multi-billion enterprise to such a bizarre failure even in understanding basic, obvious reality around them is not only self-conflicting, It's a psychiatric issue! - I think these people should be put in an institution or something

Ask me what I think of Adobe and Flash now (that's another one staring down the same yet-smoking barrel that shot M$'s foot off)

so well - computers themselves cannot be paradoxical in any way, they'd simply crash - But people, on the other hand....

[ChrisWill]

Windows 8 could be considered a paradox...

Haha, yeah unfortunately Windows 8 is not the greatest.

[Scotius]

Ah, dangit! Now i want *facepalm* emote too...In today's world it would be used a lot

@Crush. Nope, i've checked. That quote was indeed from the short story "Lion of Comarre" written by Clarke. But now that you mentioned Asimov, i think i recall something similiar from his works. Man, it's been a long time since i've read Asimov 'robotic' stories

[jebbe]

Not exactly a paradox, but:

#define true false

[ChrisWill]

That's actually a pretty good one jebbe.

[JebidiahsBigSister]

There are no such things as "programming paradoxes" I've been a professional Developer since the 80's so I know something about Computer Science . There are however some concepts which are extremely hard to prove or disprove such as the n=p problem (google it) which rely on an advanced university level understanding of information theory and is out of context to really discuss here.

What you are talking about , The Singularity, is science fiction (look up Vernor Vinge) , not science - all too often the two are confused.