The Nineteenth Byte - 2014-09-02 (page 1 of 4)

« first day (1312 days earlier) ← previous day next day → last day (3514 days later) »

New Sandboxed Posts

1:24 AM

0

A: Sandbox for Proposed Challenges

Determine if 1/x (Where x is prime) has a recurring period of x-1 This is going to get down to the nitty-gritty... Fairly straight forward stuff. Code golf a program that will accept a positive non-zero integer (x), and if x is not a prime number, report that this is not valid input with the wo...

2 hours later…

3:20 AM

Hello

3:31 AM

Hello?

3:43 AM

I'll look at your sandbox entry

I don't know how interesting that is

A similar idea is to take 3 2D points and output triangle centers

there are over a hundred well-known types of triangle centers

4:35 AM

fine.

@PeterTaylor I saw you answering some questions on Seasoned Advice

5:18 AM

@Sparr @PhiNotPi CodeBots2 spec is done. I'm writing the controller right now. It will not be sandboxed, but it is event-driven, and not in 3D.

6:15 AM

@Nathan, interesting

do you know when you'll run codebots again btw?

1 hour later…

7:23 AM

@EricTressler Yes, it's one of few handful of stacks I visit.

Martin Büttner

7:51 AM

@PhiNotPi yes

Martin Büttner

8:12 AM

@PeterTaylor I'm about to finish GEB... can you recommend any decent literature on number theory? ideally something that gives a good general overview? or should I rather just google around and see what I can find online?

@NathanMerrill the new scoreboard is crazily different

8:49 AM

@MartinBüttner The book that comes to mind is Concrete Mathematics by Graham, Knuth, and Patashnik. It's not just number theory, but it has a chapter on it, and it's a good book to have around.

9:33 AM

hi

hi

do I just post an answer to meta.codegolf.stackexchange.com/questions/1847/… ?

to add my new question?

yup, that would be the recommended way to go about it

just a second

ok thanks

1

A: How does the Sandbox work? How do I use it?

Template for Challenge Proposals Every proposal should contain several fixed pieces of information, including a title, the tags you intend to use and body of the challenge as you intend to post it on main. For a more uniform look of the sandbox (and easier orientation of reviewers), it is recomm...

here's a template

after you've posted your suggested question, a bot will post it here within 10 minutes

but your free to promote it yourself too

bugging people personally is the best way to get feedback

9:43 AM

:)

Martin Büttner

10:06 AM

@PeterTaylor Whoa, that's quite expensive. Thanks for the recommendation though, I'll keep it on my wishlist for now.

Interesting, "using built-in functions" is no longer a loophole, by a decent margin

(+33/-19)

That is interesting

maybe there's a misconception

people might assume this implies built-in functions aren't allowed at all

Martin Büttner

No I don't think so (especially if they read the description and the comments)

I think I might need to be worded more clearly as suggested in the comments

(with 33 upvotes)

I'm pretty sure some of thos 33 didn't upvote or even downvoted the answer

Martin Büttner

Well, one of the downvotes is mine. I'm with daviewales on this one. Choosing the right tool for the job is part of the golfing challenge.

damn, it's actually +35/-19 now :D

Yeah, but if there's an obscure language that happens to solve a problem that is otherwise interesting in a single command

WHOA WHOA WHOA!

Martin Büttner

10:23 AM

if you have a simple problem that is likely to be anticipated in some language, rule built-ins out in the challenge yourself (like a golf for re-implementing some date library feature, or the recent voice recognition one). if you have an advanced problem, that shouldn't usually be problematic, because it's very unlikely that some language contains that particular feature. and if it does, nice find!

whoa?

there has never been a tron challenge on ppcg?

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Can you beat the optimizer? code-challenge fastest-code This is an optimization/algorithms challenge. Given two sets of points X and Y, which are to thought of as points on a circle of circumference T, the challenge is to find the rotation which minimizes the distance from the set X to the set ...

@MartinBüttner yeah, but it's not really in the spirit of code golf, is it?

Martin Büttner

9

Q: King of the Walls

Notice This challenge has ended and will not be re-judged, but feel free to post answers and test your program against the others with the Control Program! The aim of this challenge is to make an AI to win a fight against another AI by strategically drawing a wall on a 25x25 grid to blo...

king-of-the-hill

@MartinBüttner ok

weird that the term tron was never used

Is there any way to do another tron chalenge ever again on ppcg, without changing the rules drastically

Martin Büttner

10:33 AM

probably, even just by using a continuous domain, introducing acceleration and deceleration and putting more bots in the same pit

I had been thinking of sort of a cross over between tron and vector racing

be it in discrete or continuous space

Martin Büttner

sounds interesting :)

actually, discrete would probably be better

Martin Büttner

well even discrete you get numerical issues if you start drawing walls and doing mid-move collisions.

it would probably still be easier to implement if there are only 9 possible choices per turn

that's what I was thinking

Martin Büttner

10:36 AM

feel free to reuse code from my vector racing challenge

mid move collisons shouldn't be too hard

Martin Büttner

no, not hard, but inaccurate

I'll ponder on it a bit

@MartinBüttner why is that?

Martin Büttner

because floating point

i'd just put it in a square grid though

no walls

Martin Büttner

10:38 AM

but you've got diagonal moves

X...
##.Y

here... do you hit the corner or not?

you leave lines behind, not walls

Martin Büttner

doesn't make a difference either

you've got diagonal lines and diagonal moves

intersections can be in a single point, which might not have a finite binary representation

and then a collision depends on which way you round

I see

You can do it all in rationals

Although they might get quite big

@PeterTaylor I was thinking the same

Martin Büttner

10:59 AM

@PhiNotPi any progress on Code Gofl?

11:16 AM

Is code gofl basically generalising codegolf.stackexchange.com/q/11561/194 ?

Martin Büttner

ummmm. yes exactly.

so much for novel challenge ideas...

scoring would have been shorter program's code length + levenshtein distance, which should discourage implementing both solutions and changing one character

Or, as in that case, implementing both programs with a bit of logic to work out from the input which one was wanted.

Martin Büttner

yes... I think PhiNotPi also wanted to choose more distinct programs than that

@PeterTaylor I suppose this one would be easily worked around by making both programs work on the same input.

(or one working on a superset of the input of the other)

which would still allow a maths-based program and a program taking arbitrary strings

hm, another idea to prevent implementing both programs in one is to require the two programs being written in different languages ... of course that raises questions again what constitutes a different language.

That sounds awfully like a polyglot question we saw recently.

Martin Büttner

11:32 AM

well yeah, you're right, you could still solve it with a polyglot

but if the programs are completely different, I don't think there's any benefit in including both programs one, because you'd need to switch between them, so the total code size is more than the sum of the two programs (which I think is the worst case score if you have two completely different programs and need to change the shorter into the longer)

Martin Büttner

11:53 AM

PPCG seems unusually dead this week

@MartinBüttner it really does

githubphagocyte

I don't think PPCG is dead just a little drowsy

Martin Büttner

well, I'll get some lunch now, and fiercely hope there's something to golf when I get back :D

Maybe I'll try and think of something small and simple

and hopefully interesting

Martin Büttner

btw, what do you guys think about editing a by-language scoreboard into this challenge? codegolf.stackexchange.com/q/69/8478 ... it's the plain quine challenge on PPCG, and I think the-shortest-quine-in-every-language is a much more interesting resource than the-shortest-quine-overall.

12:06 PM

I'm all for it

Martin Büttner

(I'm already looking to add answers in languages that haven't been used yet, like my Julia quine last week) ... if people saw that they could get some recognition by finding a shorter quine in their language (regardless of the GolfScript and HQ9+ submissions), this might actually get some new attention and many more interesting answers.

if no one objects, I'll do that over lunch

no objection there

I'm still wondering if a marbelous quine is doable

githubphagocyte

@MartinBüttner we might need a discussion about which answers are valid before tabulating the winners per language. There's a lot of upvoting of answers most people would consider cheating...

Martin Büttner

12:27 PM

@githubphagocyte yeah good point. but these seem to be limited to a few languages, where there are no honest contestants (at least I haven't found any, but I haven't looked in detail). also I might just include true quines.

if you guys are around, I can just post the questionable ones here and we can discuss it

githubphagocyte

I'm around.

I'm here

githubphagocyte

There are certainly enough answers to make a summary per language useful.

Martin Büttner

maybe we could even make that a policy to add per-language summaries when the number of answers on a code golf exceeds one page or so. that would make challenges with many answers more interesting in general.

githubphagocyte

To avoid questions coming up repeatedly, it might be useful once the scoreboard is in place to add a standard comment to every answer that has been excluded, so we don't get people editing the scoreboard thinking it was overlooked.

Sounds like a good habit to introduce

Martin Büttner

12:31 PM

okay, so what about GolfScript and HQ9+? they both have only one submission each

I hate HQ9+ anyway, but it wasn't a loophole back then

githubphagocyte

I know nothing about GolfScript but the "1" answer doesn't seem to be actually using golfscript, and the comments suggest it isn't a quine

What's the best place to link to as a standard definition of a quine?

Martin Büttner

wikipedia only defines it as a fixed point of a programming language

that would include all the cheat quines

githubphagocyte

@MartinBüttner do you mean the "Q" answer wasn't a loophole back then? Or is HQ9+ itself considered a loophole nowadays?

Martin Büttner

@githubphagocyte HQ9+ is

@PeterTaylor Is there a good objective definition for "true quine"?

githubphagocyte

In that case should we include it in the scoreboard with an asterisk to say this language is now considered a loophole?

That seems more useful that omitting it since it will let people know that it is no longer used

Martin Büttner

12:35 PM

something like that

yeah I wouldn't omit it for HQ9+ being a loophole

I'd only omit it for being a cheating quine

because it involves reading the source

githubphagocyte

It's interesting to describe it as a "fixed point". That makes me wonder whether there are stable and unstable quines.

Martin Büttner

most likely

most of the error quines are stable quines

githubphagocyte

ah of course

I'm now imagining an oscillating quine that gradually converges

Martin Büttner

the problem with actual quines is that you probably won't find a closed subspace of valid programs which produce each.

in almost any case, where you "almost" produce a quine's source, you're likely to generate something invalid

githubphagocyte

You mean there would always be at least some small changes which would break the program and prevent it converging?

Martin Büttner

12:39 PM

yes

I am most likely too late to the conversation about PPCG being dead this week, and I was gong to interject a Spongebob reference. "This plaque is to commemorate the brave Code Golfers who gave their lives to keep this box safe from the PopCon Menace."

githubphagocyte

So are we agreed to omit both the GS and the HQ9+ answers and add a standard comment linking to a suitably restricitive quine definition?

Martin Büttner

@githubphagocyte yes, in hopes that we can find one

(if not, the standard definitions on meta, definitely need one)

I agree on H9Q+

why exactly is GS invalid?

githubphagocyte

Even if we find a suitable definition off-site it would be handy to add it to the standard definitions too

Martin Büttner

12:42 PM

@overactor see the comments

No, Ilmari, even that is not a quine. A quine is not just a program which prints its own source, it has to have a specific structure to be a quine. – Pseudonym Feb 12 at 5:59

I wonder where he got that from

@MartinBüttner No progress on Code Gofl.

@PhiNotPi you left me last night, looking like I was talking to myself

did you see what I said about triangle centers, though?

Martin Büttner

omitting this one: codegolf.stackexchange.com/a/7347/8478

I just saw the triangle thing.

12:46 PM

Oh. I genuinely think it's interesting

something else I didn't say, because you left: If you iteratively take the triangle center, you get some cool pictures

Martin Büttner

should I exclude this one? codegolf.stackexchange.com/a/4831/8478

I.e. you take 3 points ABC, plot them and their center, and recurse on the 3 new triangles from ABC+newpoint

the nature of the diagram changes depending on the type of triangle center

@MartinBüttner no

githubphagocyte

@MartinBüttner agreed on excluding the Unix script one

@MartinBüttner because you have to exploit a language to pull it off; it isn't within the set of normal goals of a language

Martin Büttner

@EricTressler what? why?

githubphagocyte

12:49 PM

@MartinBüttner yes I'd definitely exclude the PHP one - funny but including it would require including pretty much all of them...

Martin Büttner

one thing that might help in a definition for a true quine is that removing any single character should break the quine-ness

and the empty string is not a quine

wait

githubphagocyte

Unless outputting text isn't one of the normal goals of a language, I don't see why making a quine is outside the expected capabilities of a language

Martin Büttner

that doesn't make sense

@MartinBüttner sorry, along with my speed-throttled internet, I also seem to be getting massive lag in chat. I'm occasionally getting anti-spam messages when I try to talk

@githubphagocyte I don't see why a quine should be within the capabilities of any language either. It's certainly not what any designer had in mind

githubphagocyte

@MartinBüttner the wikipedia article mentions "radiation hardened quines" meaning ones which are tolerant of any character being omitted.

@EricTressler If the designer had in mind the list of potential uses of their language, it wouldn't be much of a language

Martin Büttner

12:52 PM

@githubphagocyte hm yeah, for my definition attempt

omitting this one: codegolf.stackexchange.com/a/4826/8478 (depends on file name and reads source)

I agree that they're neat, but I also think that quineability is not important; there are probably lots of languages that are TC that don't permit quines

githubphagocyte

@MartinBüttner I think that breaking when a character is omitted is a property of a quine that doesn't necessarily follow from its definition. Might there be examples of stable quines that are still valid quines?

Martin Büttner

@githubphagocyte possibly

githubphagocyte

@EricTressler Any Turing Complete language has a notion of input and output, right?

@githubphagocyte yes, but printing a Quine is "outside the scope" in a manner of speaking

Martin Büttner

12:55 PM

what about this one? codegolf.stackexchange.com/a/2149/8478 the second one is valid and shorter than codegolf.stackexchange.com/questions/69/… ... however, it's only shorter because it uses ECMAScript 6 whereas the other one doesn't

@githubphagocyte and actually, no

@githubphagocyte Conway's life is TC

Martin Büttner

@EricTressler you still define input and output to determine the computed function

true

@overactor I agree, its crazily different. It is correct, though

Martin Büttner

should I add JavaScript and JavaScript (ES6) separately?

githubphagocyte

12:56 PM

@EricTressler I thought that saying that Conway's Game of Life is Turing Complete was just shorthand for saying that it is capable of supporting a Turing Machine?

Martin Büttner

@githubphagocyte that's the same thing, isn't it?

What's the difference?

githubphagocyte

The Turing Machine itself still has input and output glider streams?

That is the technical definition of turing-complete

it really is; it's not a gimmick, you can implement a universal turing machine in Life

but a quine is a meta-problem

githubphagocyte

@EricTressler it's already been done, so there is no argument over whether it is possible. I just still see that as a machine that takes input and produces output

12:58 PM

it's apart from the completeness of a language

????

I'm sorry, I don't understand what you're saying. My entire point was that not all languages need to admit quines

Even if they're turing-complete

« first day (1312 days earlier) ← previous day next day → last day (3514 days later) »

Transcript for

Sep '142

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile