The Nineteenth Byte

Pureferret

12:13 AM

So...this sandbox thing...I didn't realise it was a thing.

hosch250

12:37 AM

@Pureferret Yeah, it is kind of invisible.

Bye now.

mniip

4:39 AM

Aw. Rep cap, we meet again...

Quincunx

6:19 AM

Can anyone check my regex complementing skills? I say that ^([helo][world])+ becomes (^(..)*[^helo])|(^(..)*.[^world]):

1

A: Proposed questions sandbox - Mark X

Complement a Regular Expression Given a regular expression as input, output its complement. If the given regular expression matches a string, its complement should not match it, and the regex: (regex given)|(complement) should match any string. Regex for input and for output will be POSIX ERE, ...

Peter Taylor

7:35 AM

@Quincunx Without a $ anchor, ^([helo][world])+ matches iff ^[helo][world] matches, so the complement is ^($|[^helo]|[helo]$|[helo][^world])

Gareth

11:06 AM

0

Q: How awesome-ly can you print 394?

The Harry Potter series is one of the best-selling book series ever. Add to that the 8-film franchise, and it may well be the best selling series of all time. Severus Snape once ordered a class to: "Turn to page 394." Let's see who can come up with the awesomest way to display the number 394....

I don't understand how anyone can think printing a number is a challenge. Sticking a popularity-contest tag on an inane task doesn't make it any more challenging.

Jan Dvorak

This question appears to be off-topic because it is unclear to people unfamiliar to the franchise at hand. — Jan Dvorak 19 mins ago

Rusher

1:53 PM

I have a bunch of questions regarding code-golf and I'm not sure if I should put them all on meta in one post, many posts, or ask them here. Ideas?

Gareth

2:05 PM

@Rusher What kind of questions?

Rusher

Since it may be an hour before someone sees this, I'll just ask.

- Is http://www.lettercount.com/ with character count mode selected a solid tool for checking character counts?
- Is 3-5 days an acceptable time limit?
- I've been imposing input restrictions on questions and some people choose not ignore the restriction. The most refined verbage I have is "may be provided from stdin, read from an arbitrarily named file, or accepted an an argument to the program". Can that be refined further, or should that restriction simply be removed?

:-O Question vomit. Sorry!

Gareth

@Rusher Re: whitespace - if no-one says otherwise whitespace does count towards the character count. That's one of the reasons you'll find lots of answers with code all on one line - trying to eliminate the newlines. That question you link to added the extra 'whitespace won't count' rule and was exploited because of it.

Rusher

You just said two opposite statements.

Gareth

@Rusher How so?

Rusher

"If no-one says otherwise, it doesn't count."

"The question said otherwise, so it didn't count."

Gareth

2:12 PM

@Rusher If no-one says otherwise whitespace does count. :-)

Rusher

I can't read, thanks for being patient! Sorry

Gareth

@Rusher There's no rule about time limits. Some people never accept an answer because they feel it'll stop people from answering their question in the future, others accept the first answer that arrives and update the accepted answer as new shorter ones arrive. My personal opinion is that you should accept an answer, but that the time limit is up to you and whether or not you keep it up to date is also up to you.

TheConstructor

@Rusher if people provide ungolfed versions they enter the competition with an unnecessarily high number of bytes. Their problem. Of course you are allowed to give hints on how to spare some bytes but I would generally let them to the job as golfing can make the code more fragile or even break it. You are not their debugger ;-)

Gareth

@Rusher It's up to answerers to golf their code. If they don't they won't get any upvotes and they may well get downvotes. I'd just comment on their post and point out that it's code-golf and they're expected to at least try.

Rusher

Thanks again. Question 3 was only partially answered in the tag for "Code-Golf". It says input formats are specified but says nothing about input sources. I feel like it should be mentioned there.

TheConstructor

2:20 PM

@Rusher Oh and on limited inputs and other rules: tell them and neither upvote nor accept them until it is fixed. Maybe they were just concentrating on a different aspect of the task and forgot about this.

Rusher

Do you feel like it is a good restriction?

TheConstructor

It's ok. But maybe it would become even clearer if you also said what was not acceptable.

Right now I only read "you need more then some internal function to qualify" from your text

mniip

@Rusher I just use wc -c to count characters

If the question said not to count whitespace I'd add some sed magic

Rusher

Thanks all for the answers. Glad I didn't have to post all of that in five different meta questions phew

Admittedly, I could probably have avoided this by sandboxing first. Are there any plans for the sandbox to become a "feature" instead of a "question on meta"?

mniip

There's a poll on whether to move the sandbox to the main site.

Also a suggestion to link to the sandbox from the "ask question" page

Gareth

2:28 PM

I think if it was on the main site it'd get lost very quickly under all the new and updated questions.

mniip

isn't there some tag to stick a question to the top?

Gareth

There is on meta, I'm not sure if that's the case for the main site. Though I imagine they could add one.

mniip

I don't see why if it exists for meta, it wouldn't exist on main site.

Gareth

What might be good would be to disallow anyone from posting questions to the main site and only allow them to be migrated from the sandbox. Or maybe not, that would require a lot of work on the part of the moderators.

mniip

It would be great if there was a way to migrate a question to sandbox

But I don't think SE supports that

Rusher

2:36 PM

Why is the review function that rewards highly reputable users for reviewing questions not adequate for this site?

mniip

I don't think 'review' rewards anything

At least, not in reputation. There are some badges you can get from there and that's it.

Rusher

Oh, I thought reviewers got rep for reviewing. I stand corrected

I think that if Code Golf were to be held to a higher standard of question posting, a minimum rep requirement would be more streamlined than a flat "Everyone must play in the sandbox before posting."

mniip

It's already been said. PPCG contests don't exactly fit in SE's Q-A nature

Gareth

Unfortunately the horse has already bolted on that one. A lot of users have gained a lot of reputation on the back of answers to not very challenging challenges and they in turn are posting equally challenge-free questions. There's been some discussion here into what we can do to incentivise the better questions and not the trivial ones. I'm not sure we're any closer to a solution yet though. :-(

TheConstructor

Still I like it better than codegolf.com or some phpBB... ;-)

Rusher

2:52 PM

Oh, you just reminded me of another question I meant to ask earlier. Would it be a good idea to put a sort of par <1-10> on questions I post?

Even better, could the sandbox be improved if reviewers assigned a par score to the question?

That way, if I am bad at Math and Programming, I can search for easier puzzles. If I am a sadonist, I can search for par 10s.

mniip

You shouldn't exclude complex questions just because you aren't good at math.

There barely are any questions that involve high-level math. Some are just complex.
Complex, not hard.

TheConstructor

Nah, he want's to rate them so everyone finds solvable question which still demand some

Rusher

That was Math AND Programming, not Math OR Programming

Gareth

@Rusher Hmm, that is an interesting idea.

Rusher

If you want to sink trivial questions and embrace the challenging ones, assigning a difficulty level seems like the first step

TheConstructor

2:58 PM

Could be done through review and mod/high-rep only tags

Gareth

@Rusher It might be worth posting a meta question about that idea. See what other people think. I think that par-1 to par-5 or par-10 tags to rate difficulty is a pretty good idea.

Rusher

@Gareth Will do.

TheConstructor

still everyone would earn the same reputation; no matter how difficult the questions are. Also high-rep low-difficulty users could probably put a to0 high difficulty tag on questions...

But where there are humans, there will sometime go something wrong.

Rusher

Before I post... are new widgets impossible for us? Do we have to accomplish everything via the use of tags and text in the question?

Gareth

I'm not sure what abilities the mods have over us mere mortals. dmckee, gnibbler, or Chris Jester-Young might be able to tell us.

Jan Dvorak

3:06 PM

what if we ask our users to install a userscript?

TheConstructor

Propose a widget and if that is not possible go back to tags and text.

Jan Dvorak

or maybe we could allow people to edit questions in place without punishment and have askers indicate somehow when the challenge stops changing. It would definitely be less confusing for newcomers

(= ditch the sandbox idea completely and have a ready-switch instead)

Quincunx

3:32 PM

@PeterTaylor Is it right with an anchor on the original regex?

@PeterTaylor Of course, that would be ^([helo][world])+$ to (^(..)*[^helo]$)|(^(..)*.[^world]$)

Wait, I think I'd need to add |^$ to match the empty string.

Jan Dvorak

cv-pls codegolf.stackexchange.com/questions/23754/…

You have lifted a rock under which a busy ant farm toils away solving programming puzzles and participating in contests. You were probably looking for the much bigger rock covering the underground kingdom of industrious goblins who run a steampunk robot factory and spend their off hours answering random questions about programming. We call it Stack Overflow. — Jonathan Van Matre 1 min ago

2

thumb-up

mniip

4:24 PM

SO is down?

Oh it's back up.

Peter Taylor

5:01 PM

@Rusher Re time limits: I agree with Gareth but I would add one thing: in my opinion, the first question you ask yourself should be "Why does this question need a time limit?" king-of-the-hill questions benefit from a deadline because of the interaction between answers, but most questions don't.

What setting a time limit communicates to me (perhaps not to others, because I remember when they were rare) is that the person asking the question doesn't care about their own question enough to bother upvoting late answers.

@Gareth Unless there's a feature to convert an answer on meta directly into a question on main, I think that would clutter up meta with redirecting questions. If we go for a technical approach to the question quality issue, I wonder whether a good one would be for all questions on PCG main to start closed and needing 5 open votes.

hosch250

@PeterTaylor That last post is an idea.

Trouble is though, we have enough people with high rep to reopen bad questions.

It happens all the time.

Peter Taylor

Yes, I know, but even so it might be an improvement.

hosch250

Yes. We can always have a close vote button on them too, so instead of popping open/close, they just never open.

If they reach a certain number of closes, they could just disappear.

Peter Taylor

@Quincunx I think the minimal DFA for ^([helo][world])+$ has four states, of which one is accepting, so its complement has three accepting states: ^$, ^[helo]([world][helo])*$, and ^(([helo][world])*[^helo]|[helo]([world][helo])*[^world]), so one complement would be ^(([helo]([world][helo])*)?$|([helo][world])*[^helo]|[helo]([world][helo])*[^wo‌rld]). It might not be minimal, though. Your suggested complement doesn't seem to work for input h

Howard

5:19 PM

@PeterTaylor and @Gareth I don't think it is wise to implement doormen for this community. The whole idea is that it is easy to ask questions and get direct feedback. Testing questions for quality beforehand is imho counterproductive - it repels question askers if they have to wait (maybe for hours) before their contribution is accepted.

Peter Taylor

@Howard, this is PCG, not SO. There's no urgency to getting answers. And the point here is to give feedback in comments a headstart over answers which can potentially become invalid.

hosch250

I see your point, @Howard, but this community is not about feedback, it is about giving others the opportunity to solve puzzles you thought of.

Howard

Plus it is about giving other the opportunity to share their questions (without heavy barriers). I still hope that we can give the askers enough guidance such that almost all new questions become good questions.

hosch250

It is almost like making the sandbox mandatory.

hosch250

10:02 PM

Is anyone here?

@hosch250 no

OK.

:P

Are you being like Rabbit?

Doorknob

Hm, I don't understand, is that a reference to something? :P

hosch250

10:03 PM

A.A. Milne.

Winnie the Pooh.

Doorknob

oh, ok

hosch250

As we all (should) know, Windows XP is retiring on April 8th.

I was wondering if it would be appropriate to have a challenge to write a program that would notify someone who was on XP that it was time to move on?

Doorknob

that could be interesting, if done right.

hosch250

I am thinking about running it through the sandbox.

The main problem is limiting it to people who have XP.

Doorknob

Also, the trivial answer is to run ver on the command line, and inspect the results.

which isn't that interesting.

hosch250

10:10 PM

But that wouldn't prompt them to upgrade.

Would it?

No, it wouldn't.

hosch250

10:25 PM

I know, I will do it pop-contest. That will help avoid trivial solutions.

hosch250

10:36 PM

0

A: Proposed questions sandbox - Mark X

Retire Windows XP As we all know, or should know, Windows XP is being retired on April 8th, 2014. Your challenge is to write a program to notify people that their operating system is not supported anymore and is susceptible to viruses. Most interesting solution, as determined by popularity-con...

Transcript for