The Sphinx's Lair

xnor

9:00 PM

it's hard to invent a clever mapping though

user61230

"Guess the card I'm thinking of! Hint: it's red. For every wrong answer, I'll post another hint. Remember to post your answer in spoilers!"

AJ Henderson

@Emrakul oh boy

that sounds like a chat room

not a Q/A

user61230

@AJ In fairness, most of our questions aren't strictly 'questions,' either.

AJ Henderson

@Emrakul fair

xnor

i feel like the spaghetti questions were mass produced, but not neccesarry mass-producible, ironically enough

if not for the flood of derivatives, I would not have though the first couple to be a template

AJ Henderson

9:01 PM

ok, what are the "spaghetti questions"

xnor

guess-the-password questions

what's the tag?

[tag: password]

security to the party

user61230

puzzling.stackexchange.com/questions/tagged/password

user61230

thirty-six of them

user61230

I have to admit, when I first saw "[Part 26]", I thought it was a joke or something.

xnor

i had planned to satirize them that way until i saw my number was actually there

is there a more robust way to say that a question must be unique?

i don't want to ban new twists on known puzzle, but do want to ban pointless copycats

maybe one could argue that these questions are a chestnut at this point

AJ Henderson

9:04 PM

yeah, I'd argue those are a class of puzzle that is mass producible. Starting from a basic rule (the solution) and obfuscating

the solution still doesn't require particular care

xnor

i'm just having trouble expressing why these questions are low-quality in a concrete way that doesn't rule out real puzzles by being vague

puzzles where you're supposed to find a rule mapping an input to an output is a quite general class

AJ Henderson

the means of creating the problem is the key I think. If you are starting from a solution and working backwards, it likely isn't good

at best, it could have a question discussing how you solve the category of problems

but it's really just figure out the missing data

xnor

do you mean, here, arbitrarily choosing input/output pairs, then fitting a rule to them?

rather than coming up with a rule

AJ Henderson

or in the case of passwords, it is start with a rule and arbitrarily make data for it

user61230

"Make rules to fit this data"

AJ Henderson

9:07 PM

like with the animals password, the solution is how the numbers map to letters and you are then simply giving several examples of input and output

xnor

why doesn't this also rule out this though? puzzling.stackexchange.com/questions/4784/a-fragmented-puzzle

i guess you're only given inputs and not outputs, but the idea of finding a linking rule is the same

AJ Henderson

because it requires more than simply application of the rules

user61230

Uniqueness, I guess. Different logic is involved.

xnor

actually, i have a better example, one minute

AJ Henderson

if my logic is a = 1, b=2, c=3, I can choose any input for my pairs

do you have to solve a puzzle to invent the puzzle could be another way to put it

for the favorites one, you need to be able to form a word from words that fit a particular pattern

it would be even better if the words had additional relationship between them, but forming a word from words of that class is atleast reasonably challenging, even if it can be solved with a dictionary list of those types of words

it's still a puzzle (even if an easy one with proper resources)

xnor

9:10 PM

it apparently was extremely constraining, the writer has to find the few examples with a computer search

ok, i see your point about constraint

does it apply to riddles also though?

would you say finding clues and fitting it to meter/rhyme is a constraint?

AJ Henderson

I'd say that in general, if you can trivially work backward from the solution to the question, then it is mass producible

I'd have to see an example, but possibly

xnor

ok, so enforcing a ban on mass-producible puzzles sounds good

what else can be done on the quality front?

user61230

Encouraging downvotes

xnor

so that's hard

user61230

Our downvote to upvote ratios right now are... poor.

user61230

9:15 PM

Poor in a way that's shocking.

xnor

i see bad answers with many downvotes but even more upvotes

i think the issue is that site regulars are outnumbered by the masses coming in from HNQ on popular questions

AJ Henderson

@Emrakul well, not so shocking after looking at that one question you linked

but certainly problematic

user61230

@AJHenderson Fair. But, well, I'm curious, do you know what the up/down ratios are on your sites?

user61230

(Maybe post in TL?)

user61230

(I'm curious to obtain at least one point of comparison.)

xnor

9:18 PM

@joez hi

Joe Z.

@xnor: I'm here

xnor

we've been having a long discussion about what's on topic

about your post, i want to talk about the middle ground

Joe Z.

I saw the chatroom for that one question

although I didn't realize you guys were talking about it in here too

xnor

of course, the spaghetti questions are bad

but i think mathematical rigor is too high a standard

what's your opinion on things like: puzzling.stackexchange.com/questions/4784/a-fragmented-puzzle

Joe Z.

those are not language-based, they are dictionary-based

xnor

9:20 PM

so the fact that a regexp can verify it makes it OK?

Joe Z.

hold on a sec

you're right that I'm trying to rethink the "mathematical abstraction" standard

because you're also right that it feels too tight

xnor

ok, cool

i think i agree with what you're trying to express, it's just hard to pin down the boundary

Joe Z.

but in that specific case, yes, I would probably make that argument

in the absence of a better criterion, anyway

xnor

let's say that there is another solution for that mechanism

but it uses words that exist in SOWPODS or whatever and are very obscure

that a computer would find but a human would not choose over the intended solution

would that be OK?

Joe Z.

hmm

xnor

9:23 PM

alternatively, one could say that the mechanism, while verifiable, is a very human one

AJ Henderson

maybe we should post this as a puzzle and see what answers people get

xnor

post what?

AJ Henderson

it can be kind of like my non-meta meta question on moderators ;)

Joe Z.

yeah, I think the mathematical abstraction thing is more like a rule of thumb

that's probably what I'm thinking of using it as

xnor

gotcha

Joe Z.

9:24 PM

if your problem can be mathematically abstracted (including converting the problem to a computer doing something to a SOWPODS database)

then it's probably okay

definitely more likely to be okay than something that isn't

AJ Henderson

14

Q: How can I encourage more participation in a community that has been active in the past but is going through a lull in activity?

Sometimes active online communities can have a dry spell of activity. They've been active in the past, but participation has dropped off and it can become a self fulfilling prophecy that nobody participates because participation is low. How can I help boost activity in such a community and get ...

participation

that's me showing my sense of humor

xnor

for me, the condition I want to express is that once you know the solution, it's clearly the right one, not alternative solution is nearly as good, and you can either solve it or feel like you could have solved it

mathematical problems are ideal for this, as the correctness condition can be exactly specified

for riddley ones, it's very hard, but i think possible, to ensure that the intended solution stands far above the rest

Joe Z.

I think the reason that I exclude riddly ones specifically is because there's no way to tell just from looking at the question that it's supposed to have one correct answer, or that there's an objective criterion that the answers are valid

user61230

ref. multiple subjectively-correct close reasons

Joe Z.

^ exactly

xnor

9:27 PM

@Joez earlier in the discussion, i proposed a requirement that questions post an answer in spoiler

so that those who peek can see if it's bad

Joe Z.

for riddles?

user61230

That makes me a little uncomfortable, because it raises the barrier-of-entry for new users.

Joe Z.

but then why would anyone bother to actually answer it?

xnor

because riddles are fun?

Joe Z.

that's the problem I have

no, like

posting an answer in the question itself means that nobody can actually post the answer as an answer

xnor

9:28 PM

this is coming back to the question of the site goal

Joe Z.

which is further from the intention of the SE format than I'm willing to deviate

xnor

i think people will still solve questions for fun even if they could peek at the answer

is your concern that there's no way to decide what to upvote then?

or that it takes too much self-control not to look?

Joe Z.

no way to decide what to upvote, and that the correct answer as an answer would just be parroting content already in the question

xnor

ok, i see

Joe Z.

even on PCG.SE, nobody ever puts the answer in the question itself

user61230

9:31 PM

If someone posted an answer in the question, I'd edit it out and post the answer as Community Wiki.

user61230

I've actually done this once before.

xnor

@JoeZ. what do you think of that idea?

AJ Henderson

oh hey, emrakul, looks like up/down vote ratios is queryable from data.stackexchange.com

user61230

Does it include deleted votes? o.O

user61230

Actually that probably makes little to no difference.

Joe Z.

9:33 PM

that's interesting

but somehow I feel like it would encourage people to post low-quality riddles and their answers (even if the answers are community wiki), and that we'd get the same glut of them that we've been getting

like, I understand that posting your own questions and answers is acceptable on SE

xnor

i we're willing to close riddles whose answers are ambiguous, that could handle it

Joe Z.

but riddles? I'm not sure

I would say the riddles with ambiguous answers should probably be deleted instead

if we're going to mandate that riddle-askers have to post their own answers as community wiki

user61230

Creating a riddle, posting it, and then self-answering it may not be off-topic or against Stack Exchange policies, but it's still probably a waste of bandwidth.

Joe Z.

^

(i.e. that's what I'm getting at)

xnor

and you're saying it's the posting of the answer that makes that the case?

if someone else answers it, it's not a waste of bandwidth?

Joe Z.

9:38 PM

no, it's not the asking/answering itself that's a waste of bandwidth

xnor

what about PPCG then?

Joe Z.

but more the fact that asking/self-answering is supposed to be for useful information

on PCG, there are generally multiple valid answers and the accepted one is whichever one is best

AJ Henderson

@Emrakul data.stackexchange.com/security/query/248184/…

Joe Z.

I've self-answered many of my own problems on PCG

user61230

Whichever one is best on PPCG must be selected based on objective criterion.

AJ Henderson

9:39 PM

adjust the offset as desired

user61230

Those objective criterion are public knowledge

user61230

@AJ Coooool, thanks.

xnor

i see that, maybe i'm just misunderstanding "waste of bandwidth"

PPCG poses challenges that are useful to nobody

user61230

eh, 'waste of bandwidth' is a little harsh of me

xnor

nobody will use that code

AJ Henderson

9:39 PM

data.stackexchange.com is one of those fun playgrounds that probably mods make more use of than most other people, but there is still a lot you can do in there

xnor

nobody will google later and find something useful on PPCG

but people do it for fun

yes, someone asks a "question" and someone "answers" it, but there's no real use that comes from it

is that OK?

user61230

PPCG is a huge exception to the way Stack Exchange operates

Joe Z.

yeah

user61230

We should avoid basing ourselves on the way PPCG operates under all circumstances.

xnor

right, but I think that this is a good model for "challenge-answer"-type questions

Joe Z.

9:41 PM

actually, that's exactly what I'm trying to do (base us on PCG, but for regular puzzles)

xnor

and people seem to believe that posting specific challenges is on-topic

Joe Z.

but part of that is having the same quality standards that they do

xnor

given that PPCG seems to be high-quality and successful, I'm not convinced by argument that would also apply to disallowing PPCG

Joe Z.

if we're going to go for being a regular Stack Exchange website, then the Puzzle Theory advocates are right

we can't allow puzzles as challenges

the only reason I can see us allowing puzzles as challenges is specifically because we're a non-programming analogue to PCG

xnor

while i'm open to considering the option of banning challenges (as much as it pains me), PPCG is the alternative model for challenges

user61230

9:43 PM

If we look way back when, the site definition phase didn't really include many questions that were strictly puzzles.

xnor

umm, the second question

with 14 upvotes

"Given two hourglasses, one lasting 4 minutes, and one lasting 7 minutes, how do you measure 9 minutes precisely?"

user61230

Position doesn't matter too much. Look at the rest of the questions.

Joe Z.

yes, and the next one was the 10th one

not very many

I'd be okay with banning challenges as well if it'll help us regain our sense of purpose, but in that case it wouldn't be the site I signed up for

user61230

Like I say, there were a couple, but they were few and far between.

Joe Z.

and I'd probably leave

xnor

9:44 PM

@emrakul the other one (camels and bananas) was closed a duplicate for covering the same ground

Joe Z.

(well, leave being active, I mean)

user61230

Iunno. Looking back to Area 51 definition isn't necessarily the best approach, but it's important to look at what the site was intended to be.

xnor

the idea of the definition question is to intentionally make a variety of questions

Joe Z.

(I'd probably drop in once in a while to ask an offhand question like I do many other SEs)

xnor

"Because both questions are equivalent from the point of view of "which questions should be asked on the proposed website": both of them are puzzles. I think that just one question per type is enough, otherwise we could just open hundreds of questions, each of them proposing a different puzzle."

now, i'm with you that challenge questions were not intended to be such a majority

Joe Z.

9:46 PM

yes, questions about puzzles are supposed to be an important part of the site too

user61230

Questions-as-puzzles are different than asking a question about a puzzle.

user61230

"I've gotten this far, and these are my thoughts - where do I go from here?"

Joe Z.

they are

and I specifically said in that new meta question I posted that questions about puzzles aren't what I was getting at when declaring certain things to be off-topic

@emrakul: this one

but in order to cut down on the number of challenge problems being asked, there was my suggestion for what to exclude

okay, uh, brb, I'm going back home for a moment

user61230

Hmm. I think we may be manufacturing the people who think puzzles shouldn't be asked here.

xnor

manufacturing?

user61230

9:50 PM

Also, in ref. to your answer, @JoeZ: I think it's possible to disallow questions as puzzles while still allowing questions about puzzles and working through them.

xnor

manufacturing means inventing them, or producing them?

user61230

@xnor As in, who on the site thinks specific puzzles should be totally disallowed?

xnor

gilles: meta.puzzling.stackexchange.com/a/145/4551

(wow, this answer was prophetic: meta.puzzling.stackexchange.com/a/105/4551)

but, yeah, looking at that meta question, the top voted answers are "i guess, probably"

user61230

One of our most upvoted questions was simply, "why does this solution work?"

xnor

which one?

user61230

9:55 PM

40

Q: Why does this solution guarantee the puzzle is solved?

I found this puzzle online: On the top floor of a castle lives a princess. The floor has 17 bedrooms arranged in a row. Each bedroom has doors connecting to the adjoining bedrooms as well as to the outside corridor. The princess sleeps in a different bedroom each night by opening the door to ...

word-problem

xnor

oh, yes, that was a great question

it was from May 28 though, a different era

looking through your top posts, this type of thing seems to have been more common in the early days

user61230

Yeah. That's the sort of site I imagined this growing into.

user61230

Of course, when it comes to the site's ultimate future, I'm just one person. The users drive the site.

user61230

But it was different then than it is now.

xnor

i really do like your questions and answers

i'd like the site to be like that

user61230

9:58 PM

Then we need to push to get it back that way.

user61230

It's a totally feasible goal, but we have to actually work towards it.

xnor

going back to private beta was shot down, but is there some other action that would signal a reset lever?

something that would tell puzzlers who already left, "this is different, give it another chance"?

of course, posting high-quality question would help, but i'm worried that they'd get lost

and not be seen by those who'd most appreciate them

user61230

A lot of the people who don't post anymore still watch the site.

user61230

Or come back and visit occasionally.

user61230

Also, there are always new people incoming.

xnor

10:01 PM

that's true

PPCG chat periodically does puzzles in chat

user61230

10

Q: Should we have a "weekly puzzle"?

I was thinking of ways we could further engage community members, existing and new, and I had a thought: maybe we should run a 'weekly puzzle' chat event? This could be a good forum for discussion of questions that may not necessarily be on-topic on the site for being too broad, or could be a fo...

discussion chat site-events

xnor

actually, PPCG has, funnily enough, had a lot of discussion that should be in this chat room

i hadn't thought about that, but the idea of community puzzle solving is a great one and could be a big draw to the site

user61230

The problem we ran into was generating content.

xnor

people wouldn't make puzzles?

oh, shoot, i gotta run, TTYL

user61230

No worries! Seeya!

Joe Z.

10:23 PM

@Emrakul I know that; I specifically said in the meta question that questions about the disallowed puzzle types would be fine

[@ a really old message, so...]

user61230

Yeah, I just reread that. Makes sense!

xnor

11:02 PM

i have a puzzle-hunt style puzzle i've previously written I could post to test the waters on puzzles that are more involved and are targeted more at puzzle enthusiasts

is it worth a try?

A E

@Emrakul I'd have no problem volunteering a puzzle for that. If you do too and xnor does as well and we find one more person to put one in, then we could have weekly events for a month. That would be long enough to test out the concept and see how well it works.

Transcript for