# List of bounties with no deadline

This is a list of unofficial, deadline-less (hence not searchable) bounties offered by users on various challenges on the main site.

Disclaimer: There is no guarantee that the user will award the bounty for you in case you fulfill its requirement. Especially if the user isn't an active member anymore. The only guarantee is his/her written word.

# Guidelines for updating

• Answers can be commented on, allowing someone to notify the bounty's creator that the bounty may have been earned.
• If a bounty is awarded, delete the answer.

• How do you have a bounty without a deadline? – Beta Decay Jul 5 '15 at 18:11
• @BetaDecay They're unofficial bounties - if and when the feat is accomplished, a bounty will be created and the rep will be awarded. – isaacg Jul 5 '15 at 19:16
• This would be much more useful if people posted answers as answers rather than editing them into the question. That way when I see that it's changed I could just look at the answers sorted by activity (which is my default sort order on PPCG) rather than having to view the edit history of the question. – Peter Taylor Jun 30 '16 at 7:34
• There's a TNB feed for these bounties now. In order to pick your bounty up, you have to add an answer to this question. Question edits and answer edits won't be picked up by the feed. – Dennis Oct 22 '16 at 17:36
• Would it help to have a stack snippet in the question that just lists the first line of each answer as a link, rather than keep editing? – trichoplax Oct 23 '16 at 22:23
• Bounties that are only present in the meta question, and not posted as separate meta answers, leave no way to contact the bounty poster when time comes to claim the bounty. Please consider posting an answer if you have a bounty offered that doesn't yet have a meta answer. – trichoplax Oct 24 '16 at 0:49
• 100 Rep for a <100 byte answer to this question – Gryphon - Reinstate Monica Jun 14 '17 at 14:17
• @JonathanFrech Fixed. – ASCII-only Jan 3 '18 at 8:50
• @Gryphon Post the bounty in an answer, not a comment. – mbomb007 Apr 30 '18 at 21:56
• Since the problem of stuff being posted in the question wasn't fixing itself, I've taken extreme action. If people want to sort bounties by amount offered, there are plenty of stack snippets which produce a table based on the answers: one of those can be adapted. Similarly, checking whether users are "active" or "passive" can be done with the SE API. – Peter Taylor May 7 '18 at 10:34
• @PeterTaylor It should be noted that not all the bounties in the question had corresponding answers, so you may have removed some completely. – mbomb007 May 10 '18 at 20:34
• @mbomb007, I checked the first dozen manually and then I counted and reasoned that since there were considerably more answers than lines in the question I was wasting my time. – Peter Taylor May 10 '18 at 22:40
• @PeterTaylor As an example, randomra had some bounties, but this search shows he has no current answers. Maybe you could write a script? Or ping everyone who had a bounty in the question, and they can deal with their own bounties. – mbomb007 May 11 '18 at 3:25

# Python golf: Is itertools ever the answer?

I offer a 500 reputation bounty to the first user to demonstrate that his/her solution to an existing question using itertools is shorter than any solution which does not import itertools. To compete, the user must first nominate the answer (in chat or a comment). If no itertools-free Python solutions of equal or lesser length are posted after 7 days, I will grant 500 reputation for that answer.

Only pure questions with no restrictions posted before July 22, 2015 are eligible.

• What do you mean by "pure" code golf questions? – xnor Oct 8 '15 at 19:12
• @xnor code golf with normal rules; no restricted-source, time limits, different scoring systems or anything like that. – feersum Oct 9 '15 at 7:24
• I nominate my answer on brute force password guessing. – pppery Nov 14 '15 at 23:25
• @ppperry OK, I made a shorter one. – feersum Nov 15 '15 at 6:53
• I nominate this answer on ordered combinations – isaacg Nov 23 '16 at 18:44
• @isaacg Ah well technically you are supposed to nominate your own answer, but I added a 56-byte solution. – feersum Nov 25 '16 at 5:50
• @ghosts_in_the_code July 22, 2015 was the day this answer was posted. The purpose is to avoid someone crafting a challenge specifically designed to allow an answer to claim this bounty. – Martin Ender Jan 16 '17 at 8:43
• @feersum I nominate this answer of mine. – Mr. Xcoder Nov 14 '17 at 9:29
• @Mr.Xcoder Only pure code-golf questions with no restrictions posted before July 22, 2015 are eligible." feersum might still award bounties though, but I don't think you can expect one for free :) – Stewie Griffin Nov 14 '17 at 9:54

# Fitting but Unsuitable Languages: 1 x 500 points

There are some challenges for which we all want to see answers in a certain language, even if the solution is very difficult. I'll award 500 points for an answer in such a language to any.

### Gravity Simulator in Marbelous

Current shortest solution: 157 bytes in JavaScript ES6 using regex.

I don't know if this is even possible in Marbelous, a Turing-incomplete language, so I'd be very happy to see a solution.

### Minify Brainfuck in Brainfuck (Claimed by Mitch Schwartz)

I'm willing to give thousands more rep away in bounties of this type, so comment below to suggest challenges!

• Afaik, Sparr has some plans to make Marbelous TC. Is the third bounty still up if he does? – Martin Ender Jan 28 '16 at 16:16
• I'll need to know the details of the new features first, but probably. – lirtosiast Jan 28 '16 at 16:45
• I’m guessing the TI-BASIC program can’t just print out the answer? :P Also, is OS version 5.2 (84+CE only) allowed? There are some new string commands that would be helpful. – Hactar Oct 4 '16 at 12:34
• I noticed also that you were offering a bounty here. I'm not sure if you are still offering it as a bounty, but if so, adding it here might help it get noticed. – miles Oct 14 '16 at 8:46
• someone did the brainfuck one – Destructible Lemon Oct 28 '16 at 9:47
• What about trying this challenge in Cubix? Probably impossible – FlipTack Dec 25 '16 at 11:03
• Oh, never knew there was a Help! My calculator broke! challenge. At first I thought this was my Help! My calculator malfunctions! challenge that was linked.. – Kevin Cruijssen Sep 6 '18 at 13:27
• Maybe [this] answer is interesting. Brainfck to Unary-digits answered in Brainfck and Unary-digits. [this]: codegolf.stackexchange.com/questions/52712/… – Dorian Jun 4 at 7:55

# 1000 + 500 Rep for a meta answer that solves the popularity problem

Popularity contests have been discussed many times, always with much conflict over what counts as a pop con, whether they should be on topic, whether the rules need to be consistent between different pop cons, and with a frustrating lack of consensus.

Justin is offering 1000 rep in bounties, and trichoplax is offering a 500* rep bounty for a solution to this.

The bounties will be posted on a challenge of the answer author's choice if a meta answer presents a solution to this and meets the following conditions:

• 3 popularity contests appear on main following the new meta answer, each of which after 30 days from posting has net 10 upvotes and is open.

The meta answer can be attached to a new meta question or an existing meta question.

The bounty target can be an existing challenge or a future one (the bounty can be deferred if the challenge is yet to be posted). It can be any challenge type.

The meta answer can make the rules on pop cons more restrictive, less restrictive, or any other combination of changes, provided it meets the voting criteria specified above.

There can be an arbitrary delay between posting the meta answer and the arrival of the 3 popularity contests on main.

*The 500 rep bounty was originally offered to any solution to the popularity problem, including a meta answer with the required support that bans popularity contests altogether. Such an answer will still earn a 500 rep bounty from trichoplax, but the main wording of this post has been edited to reflect what is required in order to win both bounties. Aside from an outright ban, any other solution will not be eligible for the 500 rep bounty until the required popularity contests have appeared on main.

• I'd add another 1000 rep on top of this. My terms might be slightly different in determining if someone solves the problem, but overall it would be basically the same – Justin Jul 2 '17 at 0:12
• @Justin I'd be interested to hear your terms. If they are similar enough, we can match our terms to each other and offer a combined bounty. If they are not, we can still offer two similarly themed bounties - either in separate answers here or in one answer with two descriptions. The advantage of a separate answer is that it makes it easy to ping the appropriate person to claim the bounty, but since you've commented here we should both be pingable on this answer now, so I don't mind either way. – trichoplax Jul 2 '17 at 7:52
• I guess I want to reserve the right to delay the bounty until we see a few fixed popularity contests get reasonable reception. I really don't want to give out the bounty unless people can make good challenges from the new scheme (because otherwise it didn't work IMO) – Justin Jul 2 '17 at 8:01
• That sounds like an excellent condition - it is indeed effectiveness rather than popularity that I wish to reward. I'd be happy to adjust my wording to include that. To make it precise I guess we could specify something like "N popularity contests posted after the date of the answer, each open with M net votes a month after posting each question". – trichoplax Jul 2 '17 at 8:10
• So if someone suggests banning pop-cons with community support that is now ineligible? – Wheat Wizard Jul 2 '17 at 13:51
• @EinkornEnchanter Dang I didn't think about that. While that's not the solution I want, it could be a solution. It's possible that we should special case that. OTOH, there have been a lot of popcons, which shows that people like them. It might be reasonable to not allow banning them as a solution unless the person also comes up with a similar challenge style which captures the "general feeling" of popcons (by community consensus ofc) – Justin Jul 2 '17 at 15:05
• @EinkornEnchanter I edited out the mention of banning popcons because it isn't compatible with the requirement for 3 new popcons, which I agree is a better way of measuring it. I could add back in that meta consensus banning popcons will still earn my original 500 rep bounty, but I didn't give it much thought because I thought it very unlikely that such an answer would be able to gain the 90% support specified. – trichoplax Jul 2 '17 at 17:06
• @Justin just to let you know I've edited to reflect this, but I'm fairly sure this has no effect on your bounty. Feel free to edit if it does. – trichoplax Jul 2 '17 at 17:15
• Posting this to make a record of my decision. While I'd prefer a solution that didn't ban popcons, I do have to accept that banning them is a possible solution. I don't want to promise that I'd award the bounty if banning is the solution, but it's very likely that I will – Justin Jul 2 '17 at 23:54
• I believe this is impossible. :P – Christopher Jul 26 '17 at 20:25
• @Christopher I'm inclined to agree, but I'd love to see someone prove me wrong, and happy to give up the rep if they do – trichoplax Jul 26 '17 at 20:52
• I don't think this is possible - not because the problem is unsolvable, but because I don't think it's reasonable to expect the voting restrictions. Why does it matter what the votes are on any popcons that follow our proposal? – Esolanging Fruit Oct 4 '17 at 5:28
• @Challenger5 The voting requirements on main were added to ensure that the proposal is effective, rather than simply popular on meta. – trichoplax Oct 8 '17 at 1:55
• @Challenger5 The meta answer can be posted on any meta question, possibly a new question. Obviously posting it on a relevant meta question will make it more likely to gather the required votes. From the answer: "The meta answer can be attached to a new meta question or an existing meta question." – trichoplax Oct 8 '17 at 1:56
• Hoping to try this again – Christopher Jan 22 '18 at 18:29

# 1000 – 2000 rep for figuring out how to program with both halves of a Stack Cats program

This bounty needs a bit of background, so bear with me.

Stack Cats is an esolang with fairly strong constraints on its program structure. Specifically, each valid program has to have mirror symmetry, and the commands are designed such that mirroring a character results in the inverse operation. That means that the second half of the program (disregarding the character in the centre) exactly undoes what the first half did. So in a way it's on that central character to change the program state in a meaningful way, such that the undoing part actually results in the program behaviour you want.

Of course, there's only a handful of possible commands that can even be used for that single central character. For the mathematically inclined, this means that programming in Stack Cats is all about finding a conjugate relation that transforms one of 12 built-ins into the program you want to write. Nevertheless, I believe that Stack Cats is in fact Turing-complete.

While I've written non-trivial programs in Stack Cats (for example this primality test), they all rely on a certain construction that essentially works around these limitations, by skipping the entire first half of the program. This lets us put arbitrary "unbalanced" code in the second half, so we don't really have to worry about the conjugation. The basic construction looks like this:

<(...)*(...)>


The () are conditionals, which are always false in the first half and always true in the second half, so the first ... is skipped and the second ... is executed. The problem is that this is clearly inefficient for golfing: we could be doing useful stuff with both halves of the program, but we're just throwing one half away to be able to make sense of the program in the first place.

For very trivial problems, Sp3000 and I have found some programs that don't use this construction, but even those were usually found by an automated search. Here is an example of that for adding two numbers. In particular, our brute force approaches are usually nowhere near fast enough to find solutions that actually require control flow/loops. As far as I know, the only program that was ever written which uses both halves and contains a loop is this program by feersum (and given that the task was to implement an arbitrary OEIS sequence, it's likely that they just tried some programs with loops until one of them produced a reasonable OEIS sequence).

So, Sp3000 and I will together award 1000 rep to a Stack Cats answer, which:

• Contains non-trivial control flow, i.e. (...) and/or {...} which don't always do the same regardless of program input (in particular, if there's a (...) it should be executed for some inputs and skipped for others, to make sure that it isn't used to skip unused parts of the code unconditionally).
• Contains an explanation for how one can find such an answer without resorting to automated search (of course, you're perfectly welcome to make use of automated searches to get a better understanding of the structure of Stack Cats programs).
• Can't be easily outgolfed by an answer which uses the <(...)*(...)> template or which uses no control flow operations at all (to ensure that the program isn't just a linear solution like the addition program, padded with some loops which don't end up doing anything; and also to ensure that this technique actually makes use of both halves in an efficient way).

To be clear, the explanation is the most important part: this bounty isn't for finding such a solution by serendipity but for understanding how to construct such a solution.

Don't try to find loopholes in these rules — we're in no obligation to actually hand out the bounty, so if we feel like your answer is clearly not in the spirit of this bounty, we won't award it. I'm pretty sure that any answer which actually deserves the bounty, will do so quite obviously. Also, if you can solve several challenges (which are sufficiently different) with your technique — which would be a great proof that it's actually a general technique and not just a way to solve one particular task — we will award another 1000 rep to a second answer.

## 500 to 2000 rep for a quine in Hexagony

There are two ways to write a quine in Hexagony (which will probably result in very different solutions). The former is clearly the golfier version but the other one would be a lot neater. I'll give a 500 rep bounty to the first person who manages to write either of those:

• One program which is a quine when written most compactly, without whitespace and without unnecessary no-ops at the end (no-ops to pad the grid to the correct side-length are fine). (I have written one in the meantime but I'm happy to be beaten. You'll have to beat it at least by one side-length to qualify for the full bounty though.) This one has been awarded to Jo King's quine.
• I, (Jo King), will pass the 500, as well as 500 of my own rep to an even shorter quine. Again, you'll have to beat it by at least one side length to qualify (though a quine of the same side length that employs a sufficiently different algorithm may also qualify). The bounty has been claimed (in record time too!) by user202729's size 11 answer.
• Since the shortest Hexagony quine is now of side length 10, I (Jo King) will be awarding a bounty of 1000 for a shorter quine that beats it by at least 11 bytes (i.e. one row of the hexagon). If you can't quite get it that short, I will still award a 500 rep bounty for a side length 10 that uses a sufficiently novel algorithm. If you manage to golf another side length I will award 2000 rep (and my undying adoration)
• One program which is a quine when laid out nicely in a hexagon as usual, without a column of leading spaces and without any trailing spaces on the lines. There may be an optional trailing linefeed.

There should be some effort to golf the solution. And of course, the empty program is not a valid quine (I don't think there are any other possible cheats in Hexagony).

### 200 for beating the existing winner of Shortest universal maze exit string

The current winner has an exit string of length 79. A valid answer including code and a resulting exit string of length 78 or shorter will win the bounty.

The bounty will be awarded to the best answer when no better answer/edit has been posted for 7 days.

# 200-500 Rep for Japt Solutions

To celebrate hitting 20k rep, I'm offering bounties of 200 rep to anyone who posts 5 Japt solutions in a 30 day period with an additional 300 rep to anyone posting a total of 10 solutions within the same timeframe.

In order to qualify for these bounties you must either:

1. Never have posted a Japt solution before, or,
2. Have posted a maximum of 3 Japt solutions in the past with the last one having been posted 180 days or more prior to your first solution qualifying for this bounty.

solutions are exempt from those restrictions but new solutions will count towards your bounty. solutions which include Japt are also exempt, unless you were the one to add Japt to the chain, in which case requirement #2 applies.

I'll be doing my best to curtail my own Japt posts for at least the 30 days from 24/10/19.

Bounties will be suspended if my rep dips below 20k and, if, on the off-chance, this bounty proves to be popular enough, I may adjust the rep I award to avoid completely tanking it.

## In the running

• Nice. I like this trend. – Adám Feb 19 at 10:18
• This is a good idea to get people to less common languages. – Esolanging Fruit Feb 27 at 1:43
• Pretty sure I posted my 3rd Japt solution a few days ago (without knowing about the bounty). Might give this a shot? Anyways, congrats on hitting 20k! – dana Feb 27 at 9:26
• @EsolangingFruit, yeah, it's a nice way to follow on from Language of The Month. – Shaggy Feb 27 at 12:21
• @dana, I actually came up with the idea for this bounty right around the time you posted your first solution but decided to hold off until I hit 20k. Had I posted it at the time, you definitely would have been eligible so, here's what I'll do: as your first solution was posted over 30 days ago and your second 20 days ago, I'll start the counter and the clock running from your 3rd solution. So 4 more before the 26th of March and you qualify for the first bounty. Does that sound fair? – Shaggy Feb 27 at 12:26
• That's totally fine! I would just be doing it for the challenge anyways, so thanks for the offer. ASCII-Only did his pretty fast :) – dana Feb 27 at 12:40
• Just noticed you're keeping a tally here; well just did my third – Quintec Feb 27 at 13:53
• fifth rather long answer – Quintec Feb 28 at 14:29
• Here is my first – Embodiment of Ignorance Mar 3 at 2:28
• @EmbodimentofIgnorance, welcome to Japt, good to have you with us. I'll be back online tomorrow to update the list and provide help & advice. – Shaggy Mar 3 at 13:50
• 4 and 5! – dana Mar 3 at 16:16
• @dana Congrats :) – Quintec Mar 3 at 16:44
• 6, 7, 8 – Quintec Mar 3 at 16:45
• 2nd – Embodiment of Ignorance Mar 6 at 5:45
• 3rd – Embodiment of Ignorance Mar 6 at 20:50

# Build a faster Quine, 1100* reputationClaimed by ais523

A bit to my surprise my recent bounty on Fastest Mini-Flak Quine expired without anyone other than myself completing the challenge. Since I think this is a legitimately fun and interesting challenge I would like to offer further incentive for participation. Thus 100 reputation will be awarded to anyone who can beat my Quine in runtime.

100 reputation will be awarded to the first person to create a faster Quine in Mini-flak than the current answer and post the answer on the appropriate question. An additional 1000 rep will be awarded to anyone that manages to beat the current answer using a novel method (not a variant or improvement on any existing Mini-flak Quine).

If the first prize is awarded but not the second the additional 1000 reputation will be left open to anyone who can beat the winner of the first 100 with a novel method.

I, WheatWizard, reserve the right to improve my own answer at any time, thus making this bounty more difficult to attain (It really isn't that difficult to attain in the first place).

The 100 reputation bounty will be awarded by 1000000000 and the 1000 reputation bounty will be awarded by WheatWizard.

* Up to 1100 but as little as 100

• You have to double the reputation amount (up to the limit of 500) every time you re-bounty a question, so the minimum for the next one would be 200. – Dennis Jan 27 '17 at 1:59
• @Dennis Thanks I didn't know that. 1000000000 said he would supply the 100 rep bounty. Will that work? – Wheat Wizard Jan 27 '17 at 2:05
• Sure, that works. – Dennis Jan 27 '17 at 2:13
• OK, now that the reputation I got from the answer has given me sufficient rep to comment on Meta: I believe I qualify for the full 1100. You were right, it genuinely is a legitimately fun and interesting challenge. – ais523's temporary account Dec 3 '18 at 8:10

# 100 rep for Making a calculator in GNU sed

Considering it is hard to even make addition in GNU sed, It is definitely worth a bounty.

• I accept the challenge. What should be a minimum precision for division? This is not specified there. Also, do you have a time limit for a calculation? I ask because my submitted sed answer for the 1/x challenge takes an incredible amount of time to finish, but is theoretically correct. – seshoumara Apr 9 '17 at 19:12
• @seshoumara 4 under floating point/fractional output – Matthew Roh Apr 10 '17 at 7:28
• – mbomb007 Apr 30 '18 at 21:31
• @mbomb007 Re "If you didn't cheat, it would fulfill a bounty. This needs to be deleted unless you fix it.": OP are not forced to award the bounty if they think that it doesn't deserve it anyway. – user202729 May 15 '18 at 7:51
• Exactly. So if the person fixed their program, it would win the bounty. – mbomb007 May 15 '18 at 13:16
• I know I'm late, but there's a sed calculator here: sed.sourceforge.net/grabbag/scripts/dc.sed – ThePlasmaRailgun Jun 25 at 22:02

# 150 rep for a shorter C quine

Currently, the shortest known C quine (written by me) is 64 60 bytes. I will give 150 rep to anyone who tops this.

Rules:

• Must be a full program (include the main function, or otherwise define main)
• Cannot use the filename to store data
• Cannot pull data from the internet
• Cannot read its source nor its executable (thanks, Dennis)
• Must be more than 1 byte (bytecount includes flags)
• Must be system-independent (you may assume POSIX and ASCII)
• ... strictly more? Or at least? (anyway it's easily verifiable whether there is a 1-byte quine) – user202729 Jan 21 '18 at 9:51
• @user202729 do you have a proper 1-byte quine? – MD XF Jan 21 '18 at 19:36
• Does X count (must be compiled with -DX=main()puts("X");)? – NieDzejkob Feb 18 '18 at 18:48
• @NieDzejkob No, that's a standard loophole. – MD XF Feb 18 '18 at 18:53
• @MDXF so if flags used, what to output? – l4m2 May 8 '18 at 1:21
• @l4m2 If flags are used it's not just "C language", it's "C language with flags". – user202729 May 15 '18 at 7:43
• @user202729 Must be more than 1 byte (bytecount includes flags) – l4m2 May 15 '18 at 9:31

# 500 rep for a shorter payload capable V quine

Currently in V, there are two important quines:

2i2i


is the shortest known quine, but not modifiable. And

ñéÑ~"qpÿ


is the shortest known trivially modifiable quine. This is because

ñéÑ~"qp<CODE>ÿ


pretty much anything you want can go in the place of CODE, and it will still be a valid quine. You can see this on these answers:

Although these used a former version of the extensible quine, namely:

ñi<C-v>241<esc>"qpÿ


Either one could be easily modified to use this newer quine. Another good example is filling CODE with 'l's should always work. I will award this bounty to the first user to find a shorter extensible quine. For the purpose of this bounty, extensible means that any code you want to add to it will be recorded into some register and then pasted later, which is how the current solutions work (into register 'q' FWIW).

For completeness sake, I have found several 8-byte solutions:

ñéÑ~"qpÿ
ñéÑ~"qpx
ñéÑ~"qðÿ
ñéÑ~"qðx
ñ"qpÁÑ~x
ñ"qpÁÑ~ÿ


And many more trivial modifications that aren't worth posting. If you'd like more explanation on any of these, or want more examples, feel free to ping me in TNB or Vim-Golf.

# 100 bounty: Minecraft redstone prime checker

What it says on the tin.

My version of a valid prime checker uses:

• levers as input (ON = 1, OFF = 0), representing a binary number
• redstone lamp as output (ON = prime, OFF = not prime)

I offer 100 reputation to make an expandable prime checker - that means that you can expand the prime checker in any direction infinitely, with minimal redstone wiring (connect the two "modules" together using redstone) and the prime checker still works.

You are allowed to use any block except for structure blocks and command blocks.

Do I need more clarification?

• It is trivial to make a non-expanable that only takes one or two bits. (or probably any non-expandable for that matter). I don't want to tell you what to do with your rep but something tells me you might want to add a lower bound or remove the non-expanable category. – Wheat Wizard Oct 30 '16 at 23:50
• @WheatWizard That's a good point... fixing right now. – clismique Oct 31 '16 at 7:01
• Can we use command blocks? structure blocks? If not could you include a specification for what blocks we can or cannot use. – Wheat Wizard Nov 3 '16 at 15:03
• where do I post? – tuskiomi Nov 17 '16 at 21:01
• @tuskiomi Post it on "Is this number a prime?" (link here). – clismique Nov 18 '16 at 0:20
• @tuskiomi Do you have a working prime checker? – clismique Nov 24 '16 at 7:27
• @WheatWizard Fixed that. – clismique Nov 24 '16 at 7:29
• Would a 32 bit prime checker be eligible for this bounty. Or is the expandable thing something you are set on. – Rohan Jhunjhunwala Dec 24 '16 at 1:50
• @RohanJhunjhunwala I would say you only get the rep if you make it "expandable". Making it non-expandable is much easier compared to the "expandable" thing. – clismique Dec 24 '16 at 3:07
• @Qwerp-Derp so would writing a prime checking program for the Universal turing machine be considered expandable if increased input could be realized with more blocks on the tape. – Rohan Jhunjhunwala Dec 24 '16 at 13:38
• @RohanJhunjhunwala No, because that's not really "expandable" - my definition of expandable is that two or more copies of the same prime checker, with minimal redstone wiring, should act as a bigger prime checker. I should probably clarify that. – clismique Dec 24 '16 at 13:41
• IS this possible? – Rohan Jhunjhunwala Dec 24 '16 at 13:51
• @RohanJhunjhunwala That's why I'm offering a bounty for it. Besides, nothing's impossible on PPCG. I mean, look at Dennis's answers in Jelly. – clismique Dec 24 '16 at 13:52
• I have partial progress. It's definitely possible, but the amount of time required is a barrier. I'm open to someone helping me. chat.stackexchange.com/transcript/message/34627269#34627269 – mbomb007 Jan 18 '17 at 16:40
• Maybe you should make your restriction "theoretically rebuildable in Survival", that would exclude functions, cmd block minecarts, falling_block storing a cmd block, stuff in stored hotbars, etc. – Fabian Röling Oct 18 '18 at 18:16

# 500-1000 rep for further work on (-a) × (-a) = a × a

This now classic question has 3 answers at score 18. Is 17 possible?

• 500 rep will be awarded for an answer with score 17 or lower.
• 500 rep will be awarded for strong evidence that a solution is optimal, with a possible additional bonus for a proof.

These bounties can be awarded either together or separately.

# 500 rep for outscoring in Multiply with restricted operations

The current best score is 22 operations.

itx = 1/(1+a+b)     #4
nx = -1/(itx+itx)   #4
c = -( 1/(itx + itx + 1/(1+nx)) + 1/(1/(a+nx) + 1/(b+nx)) ) #14


Try it online!

I'll give 500 rep to the first better score, that is 21 operations or fewer.

Alternatively, you can get the bounty by proving that 22 is in fact optimal, say with exhaustive search. A weaker proof or lower bound can get a partial bounty.

# 900 rep for a shorter Klein Topololyglot

What it says on the tin, beat the current best and I'll give you 400 rep. (and 500 from me, Jo King)

Previous winners:

• What if we prove yours is optimal? – lirtosiast Jul 13 '17 at 22:23
• @lirtosiast Personally I doubt it is, but a proof of optimality is also worthy of the bounty. – Wheat Wizard Jul 14 '17 at 1:39
• "If a bounty is awarded delete it from its list." - Guidelines for updating. – NieDzejkob Nov 15 '17 at 15:24
• @NieDzejkob I'm not sure what you are trying to say, this bounty is not on the list, so if it were finished I would have already removed it, but the bounty is still active, so it really should be on the list. Were you trying to get me to add it to the list? I'm just not sure what you are saying. – Wheat Wizard Nov 16 '17 at 1:49
• @WheatWizard oh, I thought that "Earned by 0'" meant that this bounty is no longer active. – NieDzejkob Nov 16 '17 at 15:55
• @NieDzejkob Ok I see how this could be confusing. I've hopefully made this clearer in the most recent edit. – Wheat Wizard Nov 17 '17 at 7:06

# 50 – 400 rep for an APL answer

I will reward 50 rep if all of the following conditions are met:

1. The answer is in Dyalog APL Classic/Unicode/Extended, APL2, APL+, APLSE, GNU/APL, Sharp APL, sAPL, SAX, NARS, APLX, A+, dzaima/APL, ngn/APL, APL\iv, Watcom APL, or APL\360. Feel free to suggest additional languages along this theme.

2. The challenge had no previous answer in any of the above languages.

3. The challenge was posted before 2019.

4. The answer was posted after this bounty was set.

5. Awarding the bounty will not cause my rep to be under 10000.

The amount will be doubled once for each of the following conditions that are met:

1. The answer is well explained (as judged by me)

2. It is the poster's first APL answer, but the poster is not a "👋︎ New contributor".

3. The answer wins the challenge when posted.

## Bounty requests

To request a bounty, add an entry to the following list:

• [username](url-to-profile): [rep](url-to-answer-post), [rep](url-to-answer-post)

• Sherlock9: 200 (invalid), 200 (in progress)

•

## APL resources on Code Golf Stack Exchange

• Do not leave a comment to request a bounty! Rather, edit the Bounty requests list in the post itself. – Adám Apr 10 at 15:28

# 100 rep for a neural network (or other new thing) that writes Moby Dick

This will be awarded in Write Moby Dick, approximately when one of the following conditions is met:

• an answer is posted based on neural networks and gets a reasonable score. (It doesn't need to compete with the top answers, just do reasonably well.)

• an answer is posted that takes a different method from the existing answers, in my subjective opinion. All the current top answers use variations on a Markov model (i.e. store substrings of length n and guess the most common next letter); I'm looking for an answer that does something significantly different from this. As above, this should score reasonably well but need not compete with the top answers.

• an answer is posted that (a) uses a sophisticated method (in my subjective opinion), and (b) has a good explanation of that method, where "good explanation" means I could implement it myself without reading the code. This bounty could be awarded to an existing answer, if one is updated with a good explanation.

• Uh... -2. It seems that the community doesn't like it. – user202729 Jan 21 '18 at 9:48
• @user202729 that's... odd. But never mind, they don't need to - the bounty is between me and whoever would like to compete for it. – Nathaniel Jan 21 '18 at 10:33
• Unclear what you're asking. No objective winning criterion. – mbomb007 Apr 30 '18 at 21:34
• @mbomb007 don't need one. It's a bounty. My rep, my discretion. – Nathaniel Apr 30 '18 at 22:43
• I'm saying that's why it's downvoted. It's a really unclear bounty. Poorly worded. – mbomb007 Apr 30 '18 at 23:00
• @mbomb007 your opinion is noted, but I don't share it, and I don't care about the downvotes in the slightest. – Nathaniel May 1 '18 at 5:45
• This looks like a job for Andrej Karpathy's Magical LSTM. – Nissa Sep 10 '18 at 19:00

# 500-2000 reputation for a golfy language

This bounty will be awarded to a new golfing language (more precisely, to an answer in that language by its author to Showcase of Languages) that

• Bests all pre-2019 languages by 10% or 2 bytes (whichever is greater) in 10 pure challenges.
• The 10 challenges must be a subset of 100 consecutively posted challenges from before February 11, 2019.
• No challenge of the 10 is written by the the language's author.
• Is on TIO.run.
• Has available documentation.
• Is otherwise sufficiently non-cheaty (at my discretion). Huffman coding is okay; builtins that don't generalize are not.
• All of the above remain true for 1 week.

The bounty amount doubles for every additional 10% improvement up to a 2000 rep, so a language that improves on all pre-2019 languages by 30% in each of the 10 challenges will be awarded the maximum 2000 rep; a 15% improvement nets 700 rep, and so on.

• @A_ This is about language design, so to be eligible you must make significant contributions to the language. – lirtosiast Sep 29 at 18:27
• @A_ Consecutively posted challenges are consecutive on codegolf.stackexchange.com/questions?tab=Newest. You must be the overall winner on each challenge by 10%, so if the shortest solution in a pre-2019 language is 22 bytes, your answer should be 19.8 bytes or shorter to count as one of the ten. Tying a pre-2019 language will mean that challenge doesn't count as one of the ten, but doesn't count against you. – lirtosiast Oct 1 at 15:02

I don't like seeing good questions go unanswered, so I'm offering 200 rep each for valid answers to unanswered questions with a score of 20 15 or above older than a week. As of now (November 18th 2019), there are 3 such questions:

You can check for yourself what questions are available via search.

# 500 rep - Bubblegum Factorial

Bubblegum is designed for constant output - unless the code has a specific SHA-hash, you won't be able to write a program which takes input.

So, I will award 500 rep for a factorial program in Bubblegum. I don't really expect this to be fulfilled, as it requires major advances in cryptography to be possible.

If anyone ever completes this, post it as a submission to the Factorial challenge, where I will award the bounty.

• So if someone solves decrypting SHA-256 this they could either decide to grab 25 bitcoins or get 500 reputation points on stackexchange. – dylnan May 1 '18 at 2:49
• @dylnan ... Why not both? ;) – DLosc May 15 '18 at 2:15

# 500 reputation bounty to the first Turing completeness proof or disproof for any answer to this question based on proofs or disproofs of their respective conjectures

This is a long-term bounty which I am not making for the "lol"s.

I will create a 500 reputation bounty to the first answer to prove or disprove the Turing completeness of their program in this challenge. To celebrate the proving or disproving for any of the conjectures referenced in this challenge, I will gladly give a 500 rep bounty.

• What if one of the conjectures is proved by an outside source? Also just because a conjecture is no longer unproven does not mean that the answer is invalid. I only require answers to be reliant on conjectures unproven at the time of posting. If a proof arises later it is still a valid solution. – Wheat Wizard Feb 23 '17 at 22:34
• @WheatWizard This bounty relates to the TC proof, not the conjecture proof. They still have to prove stuff themselves. – Addison Crump Feb 23 '17 at 22:50
• @MartinEnder Oh! My apologies. This bounty is here for if any of those conjectures are proved, and if the answerer uses said proof to prove the TCness of their answer. – Addison Crump Feb 23 '17 at 23:16
• Hm, in that case, why favour proofs over disproofs? They will likely be equally impressive, no? – Martin Ender Feb 23 '17 at 23:17
• @MartinEnder Because they would then be unable to prove TCness. If they prove TCness without the conjecture, their answer is invalid anyway. – Addison Crump Feb 23 '17 at 23:19
• @VoteToClose I mean why is it more impressive to prove that the answer is TC (based on the conjecture) than proving that it isn't TC (because the conjecture turns out to be false)? – Martin Ender Feb 23 '17 at 23:20
• @MartinEnder Oh, I see your point. I still believe it to be more impressive as the answerer has to write a program to prove TC as well as apply what will likely be an incredibly difficult proof to their programming language. If you strongly suggest it, I wouldn't mind changing the scope of my bounty. – Addison Crump Feb 23 '17 at 23:23
• @VoteToClose honestly, I don't even know standard techniques for proving non-TC-ness of languages that have access to unlimited memory and non-trivial control flow. – Martin Ender Feb 23 '17 at 23:25
• @MartinEnder That's an excellent point I hadn't considered. Altering conditions. – Addison Crump Feb 23 '17 at 23:31
• @MartinEnder stimulate it with a non TC language (there are some very powerful non TC languages out there) – PyRulez Jun 18 '17 at 23:08

# 50 reputation for an Hello World version that consist of a Google’s native client executable.

These are just elf files. but with special calling conventions.

• the file should be less than 1000 bytes.
• the file should contains runnable code without modifications. So it must be a .nexe or a shared object.
• a +10 bonus if the elf file targets the armelv7 architecture (provided google chrome should load it on chrome os for arm).
• I don't believe you can offer a 60 rep bounty so I am not sure what you mean by the +10 bonus. – Wheat Wizard Mar 5 '17 at 18:57
• @WheatWizard : the upvote ? – user2284570 Mar 5 '17 at 19:47
• @user2284570 Upvotes are not for doing something extra you wanted, they're for when you think a submission is clever/valid/good/golfy. – MD XF May 28 '17 at 22:55

# 100 rep for a linear solution to this

This challenge, the perfect-palindromes challenge, has a straightforward solution, which is to partition the array into all possible, filter, and find the shortest one. However, this one is O(n!) (correct if necessary). I, Hyper Neutrino, will be awarding a 100 reputation bounty to anyone who can create a worst-case linear solution to this challenge. I'm not sure it's even possible, but I will award anyone who can do it.

If this is impossible, I will award the bounty to the first person to make the optimal-time solution and prove that it is optimal.

• O(n) worst case algorithm has been published in 2017, and it has C++ implementation here. – Bubbler Mar 14 '18 at 1:06
• @Bubbler Cool, thanks for notifying me! I can't reward a bounty until it's been posted to PPCG but the problem is copying an implementation or the algorithm from the published location doesn't really warrant a bounty... :/ – HyperNeutrino Mar 14 '18 at 2:11
• The real problem is, the whole code is 17k bytes long, even after excluding irrelevant parts, so it doesn't fit very well with code-golf. Maybe I'll port it to Python and golf under 10k(?) before posting... – Bubbler Mar 14 '18 at 4:12
• I tried porting to Python, but the code had weird off-by-2 error which depends on input, so I can't post it on the main challenge. I give up. Anyone else seeing this can take the job, or I think @HyperNeutrino can just close this since the author of the paper doesn't look like a PPCG user. – Bubbler Mar 15 '18 at 7:52

# 400 reputation for an answer to Paint Starry Night that takes top place without built-in compression

The bounty will be awarded to the first answer that takes first place (i.e. lowest score) without using an off-the-shelf compression routine. I'll award it even if it's immediately beaten by another answer. Currently the score to beat is 4159.06.

# 200 rep for an Abstract Rewriting System under 100 points

This will be awarded to the first answer to Abstract Rewriting Challenge (Cops) that scores less than 100 points and doesn't get cracked (or invalidated).

# 200 reputation for the true Alchemist

I don't know a lot about chemistry (nor alchemy), however - having written Alchemist - I'd be impressed if someone manages to use it as an actual chemistry tool:

If you manage to write a solution in Alchemist for the question Number of n-carbon alkanes, I'll reward you with a bounty of 200 reputation.

Edit: Changed the language's specs slightly to make integer output easier, proper string literals & a forgotten feature printing whether a computation was deterministic (using -d1): Try it online!

• @PeterTaylor: That was a really annoying mis-feature of the language, I changed the specs as follows: Out_"..." will print the string ... and Out_xxx will print the amount of xxx atoms. (I also changed the way string-literals get parsed, it now behaves as if it were a Haskell string-literal (eg. parsing \n instead of a real new-line or \NUL for null-bytes etc.)). (Couldn't fit the link into a message, editing it into the answer itself for the time being, until TIO pulls the latest commit) – ბიმო Jan 17 at 14:42
• @PeterTaylor: TIO.run and esolangs.org are up-to-date now. – ბიმო Jan 17 at 15:34
• @PeterTaylor: What problems? If you have ghc installed an need help, let me know. – ბიმო Jan 19 at 12:40
• @PeterTaylor: Oh wow, that's quite an old one. Instead of using Data.Map.Strict does Data.Map work? It shouldn't make a big difference. The other transformations you did seem sound. – ბიმო Jan 19 at 14:42
• That allows me to make progress; it also can't find Data.Bifunctor and suggests Data.Functor, which again allows me to make progress. Then there's a whole bunch of errors, a mixture of ambiguities between functions from Data.List and from Data.Foldable, and various methods not in scope (sortOn, pure, first, second, <*>). Meanwhile I've got my Python implementation finished and done some basic testing. I think I'm about ready to start the actual challenge. – Peter Taylor Jan 19 at 23:46
• @PeterTaylor: I'm terribly sorry about all this :( Pushed a change, that doesn't rely on Data.Bifunctor. I'm not sure about the ones with List and Foldable. I know where the issue with pure and (<*>) is, but tbh I'm not sure how to do a backwards compatible implementation just yet without rewriting lots just yet. – ბიმო Jan 19 at 23:56
• @PeterTaylor: Sounds good, will do tmw. Created a new branch here which could fix things (oldest version I should be able to test is ghc-7.8.4, but can't get it to work). Still not sure about the List/Foldable issues. – ბიმო Jan 20 at 0:33
• It is kind of interesting to see how Haskell is changing :) Btw. I can't see your changes to the esolangs entry. – ბიმო Jan 20 at 0:39
• Took me a while, but I finally got it debugged and golfed. – Peter Taylor Feb 7 at 16:14

# 500 rep for a ridiculously huge number

There are currently no valid answers to Golf a number bigger than Loader's number. This bounty will be awarded to the valid first answer that either

• Is under 256 bytes, or
• Outputs a number significantly larger than Loader's number (i.e. larger than any simple modification to loader.c).
• 256 bytes? Wow, that thing would probably win Bignum Bakeoff when it was made earlier. – Naruyoko Apr 30 at 0:15

## More unsuitable language bounties: 2 x 500

Similar in spirit to lirtosiast's bounties

• I'll add another 500 to the HexaRegex. – NieDzejkob Apr 7 '18 at 14:00

# 500 rep for a Brain-Flak answer to Balance the Brackets

The challenge is clearly brain-flak related, and should have a brain-flak answer, though this is understandably difficult.

• I was looking into this. First thought: Oh surely it couldn't be that bad. Then I realized that Damerau-Levenshtein includes swapping. Second thought: Holy ****, I don't think anyone will ever be able to do this. – DJMcMayhem Feb 1 '18 at 22:19

# 100 rep for a Haskell answer to "Make me some curry"

I'll offer a 100 rep bounty for a Haskell answer to the above mentioned challenge. I might offer more than 100 rep for an interesting answer (for example not just copying the Idris answer or a clever TH answer).

Given that tuples are tricky with Haskell and that you can't use lists (can only contain one type), feel free to define your own data-type or use an appropriate one from an existing library on hackage.

Note: This bounty is tied to the one below, if someone gets the second bounty I will delete this one!

# 200 rep for a low-level answer (machine language, C or similar)

I'd love to see an answer that's really low-level, that's the main reason I allowed taking the number of arguments as input because otherwise it seems impossible.

• For an answer in machine language, I need some way to know the type (or actually just the size) of the arguments. May I assume all arguments are of some constant size or may I get a list of argument sizes as extra argument? – wastl Sep 21 at 1:03