This challenge is not code golf. Please read the scoring before you attempt to submit an answer
You will be given a multi-line string as input. You should output the size of the largest contiguous chunk of non-space characters. Contiguous just means that the chunk is a single piece rather th...
For me I want to keep a link to the source of the challenge in the challenge itself, so it would be nice if I could predict where the page was going to be before I hit submit.
I'd like to propose self-scoring or a similar tag, but I'd like to get some feedback on whether this is useful and how the tag might work.
First the proposal:
self-scoring is a scoring criterion where the scoring mechanism and the task required are the same. This means that if you pass your pro...
"Redwolf's writing" (with 's as possessive) would be closer to y ladi srik ta Radvylf (that which is written by Redwolf)
And "Redwolf's writing" with 's meaning "is" would be ta Radvylf ni srik
Oh, ta Radvylf srik on its own (which is what Fmbalbuena asked, on second read) would mean "Redwolf is writing", although perhaps more closely "Redwolf writes"
su shilani is also part of my name though, which is what makes it "Redwolf writes programs"
CMM: would it be reasonable to answer this question with a language that I invented specifically for the purpose of doing well on that question, or would it be a loophole violation? Unlike many "invent a language for the purpose" questions, this one is resistant to "just make it solve the problem in 0 bytes" or "just add a builtin"
@ais523 I think it'd be fine, as long as it's interesting. I think it's more important to look at the reason for the loophole existing, which is to prevent trivial and boring answers.
this is one of those things that is known to be theoretically possible, but few people have the patience to actually do
(so to answer the "sock" question – I'm the same person as the ais523 on CGCC, but using a different account, this one has the rep needed to chat from Puzzling)
@ais523 Not entirely sure about the "technical rules legality", but IMO, it should be. As Redwofl said, the loopholes are basically to prevent boring "my made up language can do this trivially", not necessarily languages that are inspired by - and so naturally good at - tasks
@Fmbalbuena: note that Incident was intentionally designed as a puzzle, figuring out how to program in it is intended to be interesting, and that's why there's both a spoiler-free and a spoilers/hints version of the spec
that makes me not want to post explanations publicly, to avoid spoiling it for people who don't want it, but you should find an explanation of how to program it inside the zip file if you're interested
(however, even if you fully understand Incident, it's still a nightmare to do anything nontrivial in the language, and even some trivial things are difficult)
I used an evolutionary algorithm to help me with writing the hello world, for example…
@ais523 <first copy 1><second copy 1><first copy 2><second copy 2><third copy 2><third copy 1> When executes the first copy, jumps to third copy 1 or third copy 2?
Imagine introducing the concept of a Hello World program from that perspective: "A Hello, World! program is one of the most trivial programs to write, and can sometimes require an evolutionary algorithm to write"
@taRadvylfsriksushilani Probably, I'm just (attempting) to put it on DSO
@ais523 Do you think in floop is turing complete? If you reserve a variable as the tape pointer it's similar to Picofuck but I don't know if it still works with unbounded integers.
@emanresuA quite a large proportion of my languages nowadays are "infinite loop + if-like flow-control" (or arithmetic-like flow-control), so that construction can be TC; In Floop probably is
(this is assuming unbounded integers are supported – with bounded integers you run out of memory because there's a limit to how many array elements you can index)
@Fmbalbuena yes, whichever token is at the centre of the program (counting by tokens) produces output as a side effect when its first and third copies run
this became a running joke in the "add a language to a polyglot" challenge because people accidentally moved where the centre of the program was so often
(the token also has its normal behaviour too, the output is just an extra thing it does)
I don't think regex works at all well with Incident because there isn't a fixed list of tokens, and making pretty much any change to the source code can cause cascading side effects on how the rest of it parses
I seriously recommend using a syntax highlighter if you're trying to study programs, there's one in the Incident distribution
here's what the cat program looks like with syntax highlighting – it's substantially easier to read
@emanresuA if a string has three copies but some of them overlap each other or a different token, it gets detokenised
in the cat program, there are exactly three copies of ␠␠␠␠␠␠␠␠ (i.e. eight spaces), but they all overlap within the same run of ten spaces, so none of them count
@cairdcoinheringaahing Incident doesn't have numbers at all, there isn't a decrement command – what you saw there was an infinite loop outputting 1 bits
Make the largest number in up to 20 characters
Rules:
The program must terminate (You must prove it if it's not obvious)
The program must output the answer in base 10
in 7's encoding, 10 bytes is enough for 27 digits, so you could trivially do a literal-only answer and get quite a large number, but I think a non-literal-only answer probably does better
@Fmbalbuena not always; sometimes it's theoretically unknowable, sometimes it's difficult to work out but can eventually be proved one way or the other, sometimes it takes a bit of effort to work out, sometimes it's trivial
over the last 5 years or so the esolangs community came up with a lot of new tests for Turing-completeness, so it's often a lot faster to figure it out nowadays than it used to be
@ais523 If they're posted at the same-ish time, then use the one that gets an answer first as a dupe target for the second. If one is posted after the other, flag for a mod to close the newer one as a dupe
(or just flag for a mod, I believe they can close any questions as a dupe of any other question)
Currently whether Brain-flak with one level of loops is TC is equivalent to an open problem in mathematics. Knowing whether things are TC even things that aren't that weird can be very hard.
the language ignores the program you give it, and (by definition) outputs a mathematical fact about the nature of infinity
which is something that we believe is unprovable, you just have to assume that it's true or that it's false
this isn't a very useful language (although in another sense, it's a massively useful language as it'd tell us something that currently has to be taken on faith), and there's no obvious way to implement something like that…
indeed – this is a field where it's hard to use strong words, because at some point, it leaves the realm of computer science and enters the realm of philosophy
huh, I just found a quine question which gives a byte bonus for proper quines
I'm guessing the defaults probably don't apply to it because it predated them?
So if you ignore the stack height for a moment you can take any snippet and determine the depth it reads and the difference in output it makes on the stack.
this is very easy to do in one-loop Brain-Flak (in fact you can do it in two different ways), but some of the equivalent definitions on the Wikipedia page seem to allow it and others seem to disallow it
I think I understand – the only way you can halt is to have a 0 on top of the stack, and that must be a linear combination of values in known places – a recurrence relation can't handle all the intermediate values, but it can track the value of the linear combination so it halts in the same place
I think what's happening is that a user's posted a challenge that's unsolvable in general, but some closely related challenges are solvable, but they're aggressively refusing to acknowledge that their formulation of the problem doesn't allow their intended solution to it
I'm going to try using SEDE (or, more likely, get frustrated with SQL and download all the answers as HTML and use Node) to find possible candidates for off the charts and best explanation
This challenge is not code golf. Please read the scoring before you attempt to submit an answer
You will be given a multi-line string as input. You should output the size of the largest contiguous chunk of non-space characters. Contiguous just means that the chunk is a single piece rather th...
Make the largest number in up to 20 characters
Rules:
The program must terminate (You must prove it if it's not obvious)
The program must output the answer in base 10
