« first day (1555 days earlier)      last day (3289 days later) » 
00:00 - 16:0016:00 - 23:00

12:03 AM
84 and 75 toward Fanatic on Stack Overflow and PPCG respectively. So close I can taste it.
 
I'm at 96 toward Fanatic on PPCG.
 
NICE
Better spend the next four days watching the site, refreshing the page every ten minutes
 
just in case
 
Better safe than sorry.
 
well, I'm off to bed... task for tomorrow: find the first of those divisors which satisfies a simple inequality (but keep the overall state consistent such that the thing can be reused). this must be the most effort I put into a single golf since the Prelude quine.
 
12:08 AM
:)
Good luck!
I miss MathJax.
3
 
12:35 AM
@PhiNotPi Sorry I was gone for so long. I let the game choose random names for me.
 
What was the context, again?
 
3 hours ago, by PhiNotPi
Are all of the names (like Spawn name and Flag name), the same?
Here's a replay of my latest run in the simulation. The code used is identical to the one used in my latest run outside of the simulation.
 
Well, I don't really know what to say.
I can't see the code or anything.
 
12:52 AM
I'm posting a question on arquade as we speak. Hopefully the developer will see it.
 
Is the developer active there?
 
1:08 AM
I feel that the strategies I've made in survival mode are going to be quite different than MMO mode.
 
Same.
 
For example, there will be expansion between multiple rooms. What happens when someone conquers my room?
I feel that it might be very hard for newcomers to defend against more developed "colonies"
 
The problem is that you can only get so much energy per room.
 
There will also be a whole ton of things that now need automation.
 
You should make a module for each type.
 
1:13 AM
It's more than just the number of types. You have to automate spawn creation, choosing which sources to use, etc.
The "meta-automation"
Hmmm... we can probably do some calculations based on energy usage, without needing many details... as in... each room has a fixed rate of energy production, various items have fixed costs to build, bot parts have a fixed cost and spawning rate, etc.
 
Arrgh! I wanted to use the close/reopen review queues, and I need 1 more rep!
┻━┻ ︵ヽ(`Д´)ノ︵ ┻━┻ ┬─┬ノ( º _ ºノ)
And someone upvoted one of my questions. Thank you, anonymous internet person!
OK, this is a problem. I know I saw I had 504 rep, the upvote is still there, it even told me I had the privilege. And now it's suddenly gone.
 
Somebody un-upvoted one of your other questions, probably, bringing you back down.
 
1:28 AM
And now suddenly someone upvoted one of my answers, giving me +10 rep.
 
0
Q: Golf Your Language's Identicon

Calvin's HobbiesIdenticons are small images of geometric patterns that represent the hash value of a string. Stack Exchange uses the identicons from Gravatar as each user's default avatar image. In this challenge, we will use the Gravatar identicons as well to generate some text to golf. Challenge This stack...

 
Is anyone else slightly disturbed by the fact that in the binary representation the 0s are meant to represent white but are actually darker than the 1s?
 
It's not that weird
Or did I mean to link additive...w/e
 
I was willing to accept it. It's the same thing, really :P
 
1:54 AM
@Calvin'sHobbies I think there's a few cases where the binary version doesn't exactly line up with the original (possible due to aliasing?). An example is Java which has some squiggly lines as some of the edges.
@TheNumberOne All of the sources in survival have 3000 energy regenerated every 300 ticks.
 
I know.
 
Which give the maximum energy production of a single source to be 10 energy/tick.
 
Which means that to have more than 5 work bot parts milking it is inefficient
 
A single body part takes 3 ticks to produce, so a single spawner with a single source must produce bots with an average cost of 30 energy per part in order to continuously spawn.
(not 100% sure on the 3 ticks/part)
A group of three spawners can thus create a single part ever tick, on average.
A single part being worth 100 hits, that is the absolute maximum damage rate per tick a 1-room colony could hypothetically withstand for a non-fleeting moment of time.
That particular limitation is based on the number of spawners.
I think 100 damage per tick is actually a really practical thing for an invading army to produce.
The only defense against that is to have a massive reserve army, but there are limitations on the size of the reserve army based on aging deaths.
 
2:09 AM
Your creeps did 120 damage per tick.
 
Yeah...
 
@PhiNotPi I know, that's why the 300 allowance is there. I've changed it a bit to iron out some bumps
 
^ sounds good
So, I think there is reason to believe that it is very easy to overrun a single-room colony.
Do you know exactly how many ticks a creep lives?
I know it says 'about half an hour' somewhere.
 
If you built a bunch of walls and then maintained them with workers, you could absorb more energy.
30 min exactly
 
It's not tick-based?
 
2:15 AM
1800 ticks
tick = 1 s
 
I was thinking 1 tick is about 1 second.
 
... okay, 1800 ticks exactly
 
An 1800 tick lifespan allows for there to be a reserve army of 18000 hits for a single room. (1 body part produced per tick for the first 1800, after which they are formed and die at the same rate)
 
Ugh. It's going to take a lot of code just to check units are dead if units are given individual assignments.
Or you can just fill the room with walls.
 
I'm thinking that's going to be a popular strategy in MMO mode.
 
2:20 AM
Did they accept your application?
 
I've gotten no feedback.
Does that mean I was rejected?
(I don't know how long to wait for an answer)
 
I don't know. I didn't submit one.
 
Structures can be repaired at 20 hits per tick, but costing 2 energy.
So a single source, with all of its 10 energy / tick put towards that purpose, can repair 100 hits / tick of damage.
 
That can triple the damage absorbed.
You can also use healers that use 0 energy per hitpoint healed.
 
Ahh, I forgot about healers.
I'll have to figure out the damage rate each healer can cancel.
 
2:27 AM
180 hits / per tick / per source
 
Some of these defense methods are mutually exclusive.
 
Build a maze full of range towers.
 
And turn this into a game of tower defense?
 
I'm kinda curious as to how complicated the "winner's" script is going to be.
 
2:37 AM
Probably quite simple.
Or a neural network hosted on a different server.
What do you think is the simplest way to track dead creeps?
 
Ummmm.....
I don't know enough of the details, to be honest. Can you still access the memory of dead creeps?
I assume not?
 
Game.creeps won't have access to the creep anymore.
 
Is the situation that you have a list of tasks, have assigned a specific task to that creep, but then you need to check if it's dead to reassign the task?
 
Something like that.
 
If you store a list of creeps in a Memory object, so that the list persists between turns, would you be able to access the creep that way?
 
2:46 AM
Maybe.
 
Let's assume not, for now. You could store a list of assigned tasks in an object, and then filter the list of existing creeps for that task (you would store a copy of the task in the creep's memory as well). If the list is empty, then the creep with that task has died and you should move the task to the "unassigned" list.
 
That should work. I think I'll create a module to do that for me.
How many modules do you have?
 
Just the one from the tutorial. I'm considering doing a complete rewrite, since my current organization method can't possibly sustain itself in the long run.
And, I actually don't even use that module.
 
I have 7 modules. I'm also doing a complete rewrite.
 
To have more modules?
 
2:53 AM
To make my code more modular.
 
Just use a lot of % everywhere. :)
Modular arithmetic.
 
Haha
 
What happened here?
 
A joke.
 
It's not a real joke until at least three messages get removed in the fallout.
 
2:58 AM
Another joke ^
@PhiNotPi It would be cool if we found an undocumented way to summon Keepers to do my bidding.
 
That would be cool.
I'm not sure how much documentation there is on Keepers, anyways.
 
Good, I'm not blind.
 
There's probably a way to cheat an modify the Game objects, but hopefully they've closed off the easy ways.
 
3:36 AM
@BrainSteel I feel like that's probably how others felt yesterday while we were chatting.
 
3:57 AM
@Calvin'sHobbies Hmm, ActionScript seems a little messy
Not sure if it's because it's so pale
 
I'm pretty pale.
Perhaps I should learn ActionScript.
 
My programming experience started with ActionScript. It made Python look really fast
 
Haha
My programming experience started with VBA. I didn't understand what a string was.
 
 
2 hours later…
5:44 AM
0
Q: Can we remove the old sandbox links?

Calvin's HobbiesAt some point we decided to just have one global Sandbox. I think this is fine, I'm just concerned about all the links to old sandboxes that may confuse users, new and old alike. When I'm not logged in, and viewing the Sandbox, I can see about 7 "Related" questions that also look like sandboxes:...

 
 
2 hours later…
7:39 AM
@MartinBüttner 137
 
7:51 AM
Just within double of CJam! pumps fists
 
brainfuck seemed complex so I tried Headsecks to get a simpler image, well nope:
 
8:10 AM
@randomra The simplest BF-derivative I found was Farm
Might still be a little tricky though, considering the verbosity
 
yay, only 45 degree slopes, and simple too
 
The record for simplest overall of the ones I searched was Inline, but unfortunately it's unimplemented
 
8:22 AM
was there ever a /// (slashes) answer in PPCG? I can't find one
 
@Sp3000 what is your technique ?
are you drawing the shapes in the code ?
Also, Martin claims that he can do sub 100. I will also try drawing the shapes logic in a few hours.
 
8:45 AM
Yeah I'm drawing the shapes with inequalities
Which was the idea I was referring to in our discussion hours ago, because triangles and such can be expressed as inequalities in x and y
@randomra Martin's definitely done a few in ///, I'd link but phone atm
 
9:04 AM
@randomra I have at least two
 
I see, (it might worth adding the pronounced name for unsearchable languages)
like ><> and ///
 
hmm, true, although "slashes" is probably not so nice to search for either, but I can do that
 
there were like 45 matches for it, so not that much
 
9:21 AM
@Optimizer first pass, 92
it's still all very naive, so I feel like <75 should be possible
@Calvin'sHobbies you should probably encourage people to include the identicon in their post, not the rendered output. (so that people can see the challenge they had but without taking up so much space)
 
Nice Martin, now I don't need to try my approach :) (same sort of idea)
 
Martin/Optimizer would you mind writing a StickStack evaluator in CJam so people can check if they understand the rules/produce the correct number
I would guess it's under 15 chars :)
(I couldn't do it with no CJam experience...)
 
I'm on it
 
thanks
 
@randomra almost
ah wait, here's 13
 
9:37 AM
thanks, what the almost refers to?
 
"under 15"
 
I should find a similarly simple online interpreter for mainstream languages for the future...
hmm, S* doesn't put spaces between numbers as I expected
 
between numbers?
only if they are in an array
try ]S*
 
I see
cool
 
@Calvin'sHobbies also, the text areas in the snippet are 5 characters too narrow for me. would you mind making them a bit wider?
@Optimizer 81. I'm sure if you actually tried, you could easily go below 70.
 
9:48 AM
not with rle base
 
of course not. I'm sorry that I expected you to actually put some effort into this. :P
 
on a sunday morning ?
you bet not :P
 
are you fixing yours at least?
 
done
wow, C is sooo easy..
 
9:55 AM
meh, you are doing shape drawing.
 
although CJam is really not that bad
@Optimizer of course I am :D
 
1 hour ago, by Optimizer
Also, Martin claims that he can do sub 100. I will also try drawing the shapes logic in a few hours.
 
I never claimed I was doing anything else :P
 
I thought you would be doing something more interesting :P
because mod
you let me down
 
9:57 AM
before I go nuts, does anyone happen to know how to do CAS on an ATMEGA644p with avr-gcc? (8bit AVR microcontroller)
 
I know some of these words
 
exactly the same thought ^
 
Woo beating C. Now for another break.
 
@Sp3000 would you mind including your identicon?
 
It's just the one in the OP though, should I?
 
9:59 AM
oh, right
well in that case, it probably doesn't matter. but it's nice to see what an answer had to do when I'm looking at the answer.
 
k, I'll include it anyway then :)
 
Can anyone test my output for number of errors ?
 
why can't you?
 
it says bad input
always
and doesnt mark errors with X or report number of errors
 
it looks quite wrong anyway
the quadrants aren't mirrored but rotated
and yeah I also get 600 errors
 
10:07 AM
ohhh.. theyh are rotated...
I guess Jakube did not notice that :D
I'll fix it later
 
10:20 AM
My goal is still to somehow get within 2xCJam :D
 
which CJam answer? :P
hm, is there a 3 byte way to get either "00" or [0 0]?
(not on an empty stack)
 
Yours :P
 
I'm at 73 now
looking for one more byte, so my byte and error count are the same :D
 
Just got to take of 46 bytes then :D
 
you'd almost beat Optimizer then :D
 
10:34 AM
only 39 bytes to go :P
 
I have still no proof for the optimal Dungeon strategy. Maybe I will just ask for computing the result of the greedy algorithm... :/
not the most elegant solution though
 
what about the brute force algorithm?
for verification
 
I'm fairly certain the greedy is optimal but cant prove it
what should I use for the brute force algorithm exactly?
 
10:51 AM
at the very least, all orders of digits times all possible digging positions
that's O(n! n^n) I think
you should be able to test 5 or 6 digits with that
 
yeah but what that is good for? still not a proof
would be surprised if greedy would fail at that level
 
If it provides a counterexample it's proof for the contrary, at least
 
@randomra it's good for generating test cases
also what Sp3000 said
if it doesn't give a counter example that means it's worth pursuing a proof
 
I'm 95+% sure greedy is optimal
just hard to get a grab on a good property of the problem
even if I prove I can't expect that from the solvers
that's the main problem
and asking/doing O(n^n) solutions is not much fun
 
Do you have your greedy program up somewhere? I want to test a few things if that's okay
 
11:01 AM
O(2^n^n) is the best complexity (cough domino circuits cough)
... that can't be right, I'm sure I worked this out...
 
@Sp3000 not a nice code: ideone.com/V8pqQv difference from specs: input contains number of zeros though those can't be used
dungeon d is a list of depths
 
11:28 AM
ty
 
11:50 AM
Hm... do we ever need to dig both left/right of the deepest tunnel? Intuitively I'm thinking no...
 
12:29 PM
hm, is there a shorter way to test x > 2 || x < 1/2 than 2mLz1>?
 
@Sp3000 You almost surely never need to. This assertion seems provable. Will try to prove later.
 
@MartinBüttner z( if only integers
 
nope
I also really need 0 or 1
not just truthy/falsy
 
hello i want to know what makes this offtopic
 
12:44 PM
until two minutes ago, the lack of a winning criterion
wait
 
is it shortest bytes ?
 
is this code golf or meta golf? do we need to write an optimal algorithm, and the score is the size of our code. or does the output not be optimal and the answer is scored on the size of the output?
 
it is 2,y,+,x,+
 
yeah, I misunderstood that
 
and the code should alway output shortest string
by simplifying it
then attribiting minimal operands
attributing*
no matter the speed
 
12:48 PM
in the challenge you say that the number of operations has priority over the total string length
 
i think i misworded that :/
yes i must prioritise a criteria over another one
shortest bytes is added according to needss
it is not a must for me
if a string is longer than another one , but with less operations it is picked on
 
alright
and if there are two with the same number of operators, the shorter one is picked?
 
no need for score
yes
 
well there is a need for a score to be on topic here
 
i dont know how to settle one
 
12:53 PM
I'll try to reword it, and you let me know if that's what you want
what's the domain of x and y?
integers?
rationals?
 
yes
 
reals? complex numbers?
 
int
 
k
so / is integer division?
 
well if it is the challege would get harder
 
12:55 PM
what does 4,3,/ evaluate to?
 
let it be
1
 
im afraid noone would take part
 
are the a to z any integers or positive integers?
@AbdouAbdou have a look and let me know if that's what you want
@AbdouAbdou I suggest you try the sandbox next time, so these things can be clarified before you post the actual question and it doesn't get closed right away.
 
1:15 PM
0
A: Sandbox for Proposed Challenges

randomraBooks on a shelf code-golf sorting I have some books and a bookshelf. I would like to put as many books on the shelf as possible but I have a rule. All dimensions of the books (height, width and depth) should form a non-decreasing sequence on the shelf. This means every books has to be at leas...

 
@NewSandboxedPosts Sounds like Book Stack Sort
 
1:39 PM
@Sp3000 Similar but algorithmically different. Although my question needs a bigO constrait too as the solution is very boring otherwise (e.g. check every permutation of every subset).
There are some ok algorithms for my problem so I will try to find a bigO which let's multiple of them viable.
 
2:02 PM
Has anyone else in here been playing Screeps?
I'm trying to give a single creep multiple roles, like so:
Game.spawns.Spawn1.createCreep([CARRY,WORK,CARRY,WORK,MOVE],'worker1',{roles:['mine','transport']});
But it's definitely not working.
 
It looks like it should work.
 
No errors are thrown, but creep.roles always appears to be undefined.
 
creep.memory.roles
 
Thanks.
I knew it was probably a simple mistake.
I just rewrote all my code.
 
2:27 PM
@Sp3000 Yeah. Hopefully those are rare.
@MartinBüttner That's strange. For me they are about 5 characters too long. But sure. And I'll encourage the identicon
 
@PhiNotPi I've been meaning to. It looks interesting.
 
@MartinBüttner thanks
 
It's basically a KOTH.
so PPCG should be able to win it
 
I'll check it out in 3 weeks ;)
 
when the "early preview" starts?
 
2:40 PM
@PhiNotPi I went through the tutorial of it, but haven't really done much with it. I'll probably give it a more thorough look after finals week :D
 
@PhiNotPi when I'm done with my exams
 
@Optimizer Is it me or you you always seem to have issues with my stack snippets? :/
 
@Calvin'sHobbies dunno :/
 
@Calvin'sHobbies Stack snippet works wonderfully for me. Down to 1099 errors in ~200 bytes >.>
I'm bad at this.
 
Can we do this yet?
 
2:52 PM
Why is this a duplicate of the sandbox?
 
@Calvin'sHobbies I wanted to wait until Monday, but I don't mind either way.
 
I think Ismael is living in 2014
standard transform won't work in many browsers :D
 
3:16 PM
Can others see the sandbox links? Am I just going crazy?
 
@Calvin'sHobbies I think Doorknob is talking about SE's server-side caching.
 
Oh, then I am going crazy..
to not think of that
 
I can't see any links mentioned
 
was just gonna say
they're gone now
 
Well now I can't seem them. ok :P
@MartinBüttner Anyway, I have no scruples about nuking the list right now.
Boom
We can always rollback if absolutely necessary
 
3:35 PM
have we had 2d/3d terrain generation challenges ?
I know we've had at least 1
 
19
Q: Parallel projected voxel terrain generator

SztupYYour job is to generate a heightmap and show it as a parallel projected, voxel landscape. The rules are the following: The (heightmap of the) landscape has to be randomly generated You also have to describe how the algorithm you are using works, so everyone can learn something new here You have...

That's more about drawing the terrain than generating it realistically I guess
 
there was also a challenge about diamond square
and rcrmn has one in the sandbox
@Calvin'sHobbies edited the search suggestions a bit, and CWed the entire thing
 
do you have link to rcrmn's sandbox one ?
I am thinking of something on terms of black and white generative terrain of some kind.
 
@MartinBüttner Looks good. (My only gripe is that avatars don't show for CW.)
 
can CW question still not have non-CW answers ?
 
3:41 PM
@Optimizer If I convert every single one of them back manually.
So no.
 
and I don't see a benefit in making sandbox CW
 
@Optimizer then downvote the question and answer requesting it
the benefit is an edit threshold of 100
and proposals not counting towards the tag badge
 
but now that the list is gone, there is no need to edit it again and again
 
@Calvin'sHobbies ideally your feedback for a proposal should not depend on who posted it, I think. ;)
@Optimizer the answers.
 
@MartinBüttner Yeah yeah
 
3:43 PM
oh. you can answer CW questions only if you have 100 rep ?
how is that a benefit .. ? New users would not be able to use sandbox...
 
no
you can edit CW answers with 100 rep
 
and probably you can comment before that, right ? (let me check the rep guide)
ok. comment is 50
 
@Optimizer This on a very basic scale would be neat: freespace.virgin.net/hugo.elias/models/m_landsp.htm
 
@Calvin'sHobbies nice!. For my challenge, I was thinking a different theme. This planet one can serve as a separate challenge all together. Feel free to post it if you want.
Or I can too, if you don't want :P
 
Go ahead if you want, I'm not planning something for it
 
00:00 - 16:0016:00 - 23:00

« first day (1555 days earlier)      last day (3289 days later) »