« first day (1210 days earlier)      last day (3647 days later) » 

1:25 AM
Doorknob has made a change to the feeds posted into this room
 
I figured with all the fancy feed stuff going on in the other beta sites' chatrooms, we might as well have one here.
 
Well that was abrupt.
 
5 messages moved to Trashcan
yeah, that happens when new feeds are added. :P
(the feeds cache every question in the last 24 hours, IIRC)
 
Hmm. So does that mean every 24 hours it will just dump a backlog, or does it spit them out like the chatbot does?
 
No, it spits things out every 5 minutes or so.
 
1:32 AM
Thank god(s) (or your approximation or lack thereof).
 
(A side effect of RSS feeds is that the startup is... less than optimal. :P)
 
@TheDoctor do you play tagpro?
 
1:47 AM
Umm....
no
 
darn
i was playing a TheDoctor just now
not an uncommon name but it would have been cool
 
Currently IRCing
Mutiny on a dislikeable channel
 
 
1 hour later…
vzn
3:02 AM
hi dudes.
hows the motley crew
any machine-learning fans here? see no tag for that
kind of a hole it seems?
about the only ref on the site seems to be this question, which forbids it, huh
1
Q: The smallest self-improving program

vszThe goal is to write a program that learns from its own mistakes, and improves itself based on it. This means, that later attempts at the problem should be either faster or more accurate, than earlier ones. To deny trivial solutions, here is a set of rules (they are rather obvious): you can c...

lots of ML questions on other se sites, several have tags.
* eg on [stackoverflow.se]
 
Hi
Sorry, but that seems broad
 
vzn
broad? ML is indeed a very broad field that mixes with CS/programming. & so far apparently nearly zilch representation on this site. oops!
 
upvote 4 you
 
vzn
what?
anyway amazing how many sites its on... also
* compsci
* cstheory
* stats
* scicomp
* programmers
* quant
anyway maybe someone might be interested in the following which imho is way cool.... ping me for more info
kaggle competition $13K prizes
 
3:28 AM
 
 
1 hour later…
4:43 AM
This meta post almost seems reasonable as
8
Q: Can I compete in my own challenge?

ShadeWhat are the rules/community consensus about participating in your own challenge? Does it differ for each of code-golf, popularity-contest and king-of-the-hill? The person that asks the question isn't the one who picks a winner so...

I'm reluctant to delete my comments in this discussion because there are other comments that make no sense without mine. @Doorknob perhaps you can nuke?
I don't get why I keep getting Error: Could not find or load main class SwapBot. I input the string java SwapBot. I am so confused; it worked before... — Quincunx May 15 at 19:28
 
5:21 AM
I was trying to stay away from Codegolf but I finally gave in :)
0
A: Write a hello world GUI program that closes itself after three seconds

AsheeshRShell and gedit - 27 characters timeout 3 gedit Hello World timeout utility runs a command for the duration specified. It ensures that gedit runs for 3 seconds, assuming minimal startup time. Any editor can be used in place of gedit. If a shorter named editor is used like gvim, the leng...

 
@AsheeshR Welcome!
 
 
1 hour later…
6:37 AM
@Quincunx Quick question: When does a code-golf end? Is there a set duration?
 
@AsheeshR usually we check back for further golfed answers. If one beats the accepted, we change the green checkmark.
However, we usually accept after ~1 week or no activity
 
Ok. I was thinking it would be around a week.
 
But if the question is still very active, we usually wait longer
This is for every challenge type
 
7:07 AM
0
Q: Suggested Edits: Reject Golfing?

QuincunxEvery now and then, I see an edit where somebody does some golfing (most recent). From what I can tell, our standard is to reject those edits as invalid and post a comment saying what was suggested, because we feel that golfing tips should be comments. And I have always done so (after I learned)....

 
What is this user doing:
@NewMetaPosts Oh. Didn't remember that you post. Now I can't delete mine.
 
7:40 AM
@Quincunx Looks like nothing so far.
 
8:22 AM
@Geobits yay, I won the first bout, hurray for UntangleBot
By the way, how long did the tourney take ? I tried to pit all entries, too, but had to give up halfway through, those one-minute bots were too annoying
 
 
4 hours later…
12:14 PM
@Quincunx Done
@AsheeshR you will NEVER leave you are TRAPPED here FOREVER we will-- err, I mean, welcome! :P
 
12:48 PM
It's not even clever! Or creative! It's a slightly unusual source of text and some string manipulation
 
@undergroundmonorail That's pretty much how I feel about mine at 55.
 
Yours was actually creative, at least. I laughed when I saw it.
Also, I don't know if you noticed, but I misread it the first time and made a dumb comment :P
 
No, didn't notice.
 
I thought it was randomly choosing words and you were arguing "Well, there are words in there that could be an insult!" so I commented "I got You suck at random, haha" as if that was an unusual thing.
 
If it was randomly picking them and got that, I'd say you don't suck at random and should probably play the lottery :)
 
12:52 PM
:)
 
1:08 PM
Looks like we have exactly 1999 questions right now. Anyone up for writing the 2000th? :P
 
Anyone here know anything about search algorithms? I'm trying to use Dijkstra to find the minimal path to one of any number of exit points. If I break the first time I hit a target, is that one guaranteed to be the one with the shortest path, or have I only found the shortest path to that exit?
I think it's guaranteed to be the shortest path but I'm not certain.
 
@undergroundmonorail Any early answer on a question which hits the "network hot questions" list gets upvoted a ton. That's why some people managed to build massive reputations on code trolling.
 
@PeterTaylor True.
 
@undergroundmonorail It's guaranteed to be shortest.
 
@PeterTaylor Thanks :)
 
1:10 PM
Proof: consider a modified graph to which you add a single exit node, and an edge of weight 0 from each of the exit nodes of the unmodified graph.
 
It would see that one first, follow it and break out. Right.
Thank you <3
 
@Doorknob I literally just opened the chat to see if someone had already posted that comment :D
 
@m.buettner :D Maybe we should do something to celebrate. :P
I'll think about ideas at school! (which is where I have to go now)
 
like what? ^^
right
 
I'm at school now!
Figuring out really clever stuff for Project Euler problems sucks because then I can't talk about them :c
 
1:25 PM
:D
you can talk about it in their forums
 
Yeah, that's one thing :P
 
2:00 PM
@Doorknob I volunteer as tribute
I'm about to start sandboxing a ton of crap lol
 
^ said the cat after a large meal
4
 
lol
 
2:23 PM
This is what happens when I try to grind out a challenge. Out comes a turd.
Crap I forgot the sample input
Fixed
Thoughts?
My thought is that the sandbox is a good place for turds.
But maybe you guys can improve the challenge. I'm all for playing together in the sandbox
 
@Rusher can we assume that no letter will be invalid? (i.e. all of them are found on one of the lists)
 
Yes. Each of them corresponds to one of the items in the list
I guess I should add bounds on the number of letters so that you know how large of an input your program has to handle
Because a series of 10 ?????? could get very large very fast
 
2:41 PM
do we need to handle unsolvable input (by printing 0)? e.g. if no combination gives a vowel or if all possible combinations produce consecutive vowels?
 
I guess the number of words would be zero in that case. I should explicitly state that
 
in any case, 1 or 2 more samples could be useful. such that they cover all the necessary logical deductions to fully solve the problem.
 
Made edits. Thanks @m.buettner
At most, you should only have to consider 2^30 possible combinations if I max out the number of ???s
This would have been a better fastest code challenge...
If you allow for more code, you can easily optimize the search based on the rules given
All my creative flairs are gone, and I'm not super happy about that one
 
well it's still in the sandbox, you can change it to fastest code any time ;)
code golf could be interesting, too, because I'd like to see how people compress the tables for valid letters
 
3:03 PM
Oh yea... I forgot about that
Well now that you mentioned table compression I am a little happier about the challenge
 
does the people here help others with fixing codes. I mean, making codes better, and faster.
 
@blackbee if you're talking about golfed code, yes. if you're talking about production code, then no, and you're looking for codereview.stackexchange.com
 
what is golfed code?
 
code that attempts to solve a problem in the fewest amount of characters
(for recreational purposes only)
 
o
 
3:09 PM
I'm pretty sure you're looking for codereview.SE
 
"taking code that works and making it better" is pretty much their m.o.
 
codereview.stackexchange.com/questions/51501 <--i did post it, they re-edited it, made the UI look of question better. But then there are no fruitful answers except one hint (using lookup table)
 
it's only been 3 hours
give it some time
 
@Rusher I can get a list called v of vowels and a list called c of consonants in 123 bytes. That's my first shot with this method, I could probably get it down by rearranging the order of the letters in the lists (which doesn't matter at all anyway)
 
@blackbee Change your username, come back here, strip your problem down the the bare bones, make up a story about dragons and knights and princesses, and then post your problem as a challenge
 
3:12 PM
haha
 
ya, actually you are right. i am bit impatient, i have been speeding nearly more than 3 days pondering over that.
 
You speed a lot. You're gonna get a ticket
 
Change my username? suggest one
 
it was a joke, don't actually do it :P
 
@blackbee I suggest "One"
 
3:14 PM
:D
RaOne?
 
You know what, we need more diversity
I suggest Raul, Binh, or Latisha
 
www.imdb.com/title/tt1562871
Latisha .. dats a female
 
Well I'll be honest, I started typing Trayvon and was like shit I can't say that
All I know is that we have a John, a Peter, and a Matthew in here. All white folks from the Bible. We need some diversity, and we need a female, so Latisha is actually a decent fit
I also think Lucy is kinda cute. I love Lucy.
 
As long as it's not Chris. Everybody hates Chris.
 
3:38 PM
pinterest.com/pin/349029039842144673 i honestly don't understand the second step
 
Looks like they assume you know what a square base is.
 
Papa Johns pizza is 50% off for the next three days
 
Yea but it's Papa John's.... that's barely pizza :p
 
Ok but people actually eat Hot Pockets and Papa Johns is at least three levels higher than that
You have Hot Pockets < Frozen Pizza < Domino's Pizza < Papa John's pizza
 
True, but when I microwave something, I expect it to not be great. That's like comparing a cup-o-ramen to a bowl of restaurant-style ramen.
The only real problem I have with Papa John's is the sauce, to be honest. It just tastes odd to me.
 
3:50 PM
It's kind of sweet, and so is their bread
Balances well with the yellow pepper that is kinda sour
And the garlic butter that is salty
 
Yea, and I guess that might be okay on something like a Hawaiian, where it's sweet anyway.
Recently I've been liking spinach/feta on my pizzas. Sweet wouldn't work so well there, I'd think.
 
We're twinkies! I always get Spinach and feta too!
It was meant to be
 
Oh, you forgot CiCi's in your rankings above. It should either be above or below frozen, not sure which.
 
4:15 PM
I think it fits inside of frozen. Some Frozen is better, and some is worse
Honestly I think the 99 cent Totino's pizzas are better than CiCis so maybe it is strictly worse
I might need to get community concensus on meta
 
You're probably right. Totino's makes a good snack pizza.
 
@Rusher Hey! I'm not that old.
 
Oh, you're that Peter? Never knew his last name.
 
Do we have any other Biblical names in here?
I counted Matthew John and Peter
 
I think that Peter's "last name" would have been Bar Jonah.
 
4:23 PM
Oddly enough, I work for the first book of the Bible
Mom named her company after a concatenation of two other books
 
(Whether he was "white" or not is a discussion which it might be best to skip)
 
@PeterTaylor Probably, but since the name "Peter" was given to him later, I figured he might have changed it in the meantime.
 
I thought his name was Simon Peter
Is that a different Peter?
His ethnicity was Middle Eastern nndb.com/people/237/000094952
 
@Rusher not biblical but I can offer Martin as a saint's name :P
 
I thought you were matthew lol
 
4:26 PM
He was born Simon (or the language's equivalent). "Peter" is a masculine form of "petra", meaning rock. It was given to him by Jesus, along with the whole "upon this rock I shall build my church" bit. (according to the New Testament, of course)
 
I wonder how you jumped to that assumption :D
 
@m.buettner Saw your name on some voting spreadsheet google doc and thought it said matthew
I think it was the code trolling one
 
ah okay
 
Also, I'd say Malachi fits squarely into the biblical names category.
 
I forgot my list of books so I forgot Malachi was even a book lol
 
4:31 PM
seems unlikely for a non-native speaker though ;) ... then again some people do anglicise their names in English-speaking contexts (e.g. Matthäus might be a bit awkward)
 
I like the Spanish version of my name way better than the English
Juan
Makes me feel like a winner, because I Juan.
 
^^
My name usually remains the same in other languages, except the odd Marten or Martyn (I've never come across any of those instances in real life, although I know they exist - I've only found the former in Stephen King's Dark Tower and the latter in A Song of Ice and Fire :D).
 
@m.buettner Fun fact -> You're Marton in Hungary
 
oh I've haven't seen that one before
 
4:52 PM
UGHHHHHHHHH pastebin.com/sy23j3Vd because sometimes you just need to rant to somebody
 
We have 2,000 challenges now!
 
^^
I have to give sample output later. Out for now ><
 
@Geobits for what range of θ do you want the output?
not all r/R/d pairs produce closed curves, do they?
oh they are all integers. I guess they do, then.
 
Right, I think it will for anything rational.
 
5:05 PM
yeah probably
"no hardcoded input". can I expect them to be stored in fixed variable names though?
 
If you mean R=10, then no. Otherwise I'm not sure I understand the question. They should be able to be input in some form without changing the source.
 
What does "other text-based output" mean?
SVG? PNM?
 
No, I meant if someone decided to get smart and use non-ascii characters to make a textual output.
 
@Geobits okay, I'll just make it a function then
(I meant simply acting on globals)
I think they will generally not actually be closed curves, due to floating point inaccuracies of cos and sin.
 
@m.buettner With a bit of care they can be done entirely in rationals
I think
 
5:15 PM
Huh, really? I'll have to think about that.
 
Yes. Change of variables: x = (R - r) cos (rt) + d cos((R - r)t). Pick t such that cos t is rational, and it's doable entirely in rationals.
 
This page goes into that a bit.
 
(Sorry, substitute nt for t above, where n is the step. Usual problems of fixed-step curve drawing apply)
 
Hm, I don't think I really see how that helps, but thanks for the link, I was trying to figure out the minimum θ necessary.
 
5:32 PM
@luserdroog We deleted all 7 posts.
 
@m.buettner Down near the bottom it shows that a = r / (gcd(R-r,r)) and minimum θ = 2πa.
 
@Geobits yes, that's what I was referring to ^^
 
Oh, I thought that was part of "I don't see how that helps" :)
 
oh yeah, I see how that was ambiguous
 
 
2 hours later…
7:50 PM
After collecting some ideas on KotHs a few days ago I started working on one concept, and would like to gauge how interesting and feasible you guys would find this. The current idea would definitely make it one of the more complicated KotHs (maybe similar to Geobits' Hunger Gaming). In particular, I was trying to make use of the "multiple bots/instances that cooperate" concept. If you think this has potential, I'll write up a draft of the rules and put it up for discussion in the sandbox. [ctd.]
2
For starters, I was trying to take a different approach towards KotH. Most (all?) of them currently simulate the "arena" in fixed timesteps - i.e. simultaneous turn-based systems. I thought we could try something more akin to Final Fantasy X's Conditional Turn-Based Battle system: different actions take a different amount of time, and it's simply your turn again when that time has elapsed. Of course, hangtime, would be skipped by the simulation.
Secondly, most KotHs seem to use a grid or Euclidean topology. To bring some change to that, the stage would be the solar system (all large-ish solid bodies in it, actually... those are between 50-90 depending on which ones I add). So the map would be the complete graph of those with unique (and changing) travel times between each pair.
Thirdly, I'm not sure about the combat system yet, but I would like to use Risk's system. It's not too complicated and is a nice way to give defenders a slight advantage in combat.
So these are the ingredients. The goal would be to conquer as much of the solar system as possible. You've got a mother ship that can build/take over bases and a bunch of fighter ships that actually do the purging of enemy units (here is where the communication comes into play). The mother ship is slow, so you'll also want to use the fighters as scouts to see what's going on on other celestial bodies. This is where I might be able to simplify things by scrapping the mother ship.
Otherwise, I'd provide between one and three default implementations for fighters, so people who don't have the time or patience to write two bots could just write a mother ship that commands those.
(sorry for the wall of text here in chat, but I figured it's too vague for the sandbox like this, and just wanted to get some general feedback before writing everything up)
 
8:06 PM
If all you did was build a simulator of Risk I would spend a day writing a submission for it.
 
Actually a simple Risk KotH is on my list as well (and maybe I should start with that). I just realised that its combat system would be very well suited to the above concept as well.
 
a solar system just equals a new map right? You define what is connected
And you define what groups of areas give bonuses for holding
 
basically... it's a complete graph, but with changing weights (the control program would simulate the orbits in a simplified way)
 
I'm taking that mother fucking Australia and being a pussy for the rest of the game. Deal with it.
 
:D
I'd probably make the bonuses the same for all areas. Like 10 fighters per base or so. The strategic advantages would come from the positions in the solar system, I think.
 
8:09 PM
I guess you have to add areas for each new player
Something like n-1 areas for n players?
 
I think there are enough celestial bodies in the solar system, so that we won't run out of space unless I get more submissions than your wolf challenge did. I intend the solar system to start with unoccupied bodies.
 
It looks like someone wrote a thesis on AI for Risk if anyone is interested. Here it is.
 
Germans... :D
bookmarked!
 
@m.buettner I can't figure out which of the above messages to star so others will see the outline, but I say go for it. I think a sketch (at least) of the spec will help in determining what comments/criticisms to give.
 
@Geobits star the first one, it ends with [ctd.] :D
oh right, the transcript doesn't show that much
well thanks anyway!
 
8:19 PM
It also gave me an idea for a challenge (not related, just tangent thoughts), so thanks to you, too!
 
Ha, you're welcome, I'm curious what you'll make of it ;)
 
It's really not related. You know how your mind goes off on weird paths? Mine ended up on a space/ocean-based trade sim. Buy low, sell high, etc. Still thinking on it.
 
Ah yeah, I know what you mean. Still interesting though!
@Geobits would you also think that a plain Risk KotH might be interesting regardless of the above concept?
 
I think it would gather more players, just because it's familiar. I think the other sounds more fun, but may have a higher barrier to entry (would need the spec to say for sure).
 
Yeah, I agree. I'm just wondering whether it might be worth doing both (with some time in between). I don't think the combat system would be the defining thing of the challenge, because the solar system idea would introduce a strong exploration component (you could only see the state of planets you currently have units on).
Also Risk itself has additional concepts like cards and dominated continents and different phases of a turn (and in general turn-based gameplay) which I wouldn't have in the solar system one. So imo they'd probably be both worthwhile in their own right. I'm not even sure which one would be harder to implement (both in termsof the control program and the bots).
Well I guess I'll get down to writing up a draft of the rules for mine then.
(I'm still happy to hear other opinions in the meantime!)
 
8:33 PM
oh hello random people on the internet acts surprised
 
Hello Kyle Kanos
I feel so small in this chat filled with people like Kyle Kanos who are pursuing their PhD in Physics specializing in Computational Astrophysics
That's why I make the challenges, attempt to solve the challenges, fail at solving the challenges, and make more challenges.
3
 
so i have an hour and 20 minutes until forced bedtime, gonna work on ze koth
 
did someone say physics? :)
 
@Rusher i'd say the same but this is going to be the first challenge i'm posting ^^'
 
@Trimsty Are you under martial law?
 
8:47 PM
@Rusher Parents. :P
<i believe that's a yes>
 
If you are programming, you do not need to have a curfew.
 
@Rusher +1, but they don't let me do anything :c there used to be a deal that if I got up at 7 AM, I wouldn't have any enforced bedtime
that obviously didn't work, so
 
Hah. The deals. They get rather complex up until you have a huge fight and then everyone calms down and dad gets manly and says, "Son I think you need more freedom, but you also need a job."
 
I'm taking a different strategy this time--
 
Then you get all the freedom you ever didn't want.
 
8:50 PM
@Rusher XD
okay, so i finished the first phase of development; made a folder called "thingy" on my desktop
 
9:09 PM
Hi @Rusher
But I'm actually heading out, gotta get home :(
Note that that's not a frowny face because I don't like home
 
I know. You miss me already and you are sad.
 
Maybe I'll pop in on Tuesday. Memorial day weekend and all
 

« first day (1210 days earlier)      last day (3647 days later) »