# What does "unclear" mean?

This topic was closed as unclear. However, all asked problem in the comment that matter seem answered in the first version statement.

1. The result should be constant when n exceed a constant.
2. Decimal and unary(use the output.length) output allowed
3. No infinity output, but infinite loop allowed if the loop doesn't generate any output, or the infinite output is after a decimal point.
4. You can assume sizeof(a number type) be its size on some existed env or linear to 2^n with an existed rate. That means, long long in C++ can be 4 bytes(16-b env), 8 bytes(32-b env), 2^n bytes(64-b env), 2*2^n bytes(32-b env). However, if char* is 4-byte or 8-byte, it can only handle 4GB memory or 2^64 byte memory, and the result won't be that large.
5. For float(and double), you can define your custom IEEE bit arrange, sizeof(float)=2^n, sizeof(double)=2^(n+1)
6. Unlike other models, with n treated as a large enough input, this model can solve something like the busy beaver. Therefore, you may want to go further.
a. I only answer "What the problem mean", not "Why ask this problem"(The latter usually don't apply in PCCG)


It's very unclear to me what the challenge is actually about. Given a 2^n-bit x86: Is this an assumption about computer architecture? Some kind of input? What is n? output the largest result when n is large enough: Large enough for what? [a]
A busy beaver maximises some property of the computation model (e.g. execution time, memory usage, output length) subject to the constraint that it must halt. From Note 3 I don't even think it is a busy beaver (because infinite loop is allowed). And when n is large enough, does this imply unlimited precision? [3&6]
What do you want to say? I don't understand. / What about the program unsigned n=1,p=0;while(n>p){++n;++p;}print(p); which prints the largest unsigned value?[1]
So what's the difference between this and the other challenge? Any program from the other challenge can be ported to this challenge and remains competitive.[6]


This one was closed(later reopened) also as unclear, for they didn't know the statement's truth value is unknown. However, the first version directly says

you shouldn't just say "one of E x x=x and A x x>x match"; but if you proved one, you are allowed to use that XD

Therefore "unclear" seem doesn't mean "don't know what the question mean, and thus unable to answer". So what does it mean?

• It seems that you are not a native English speaker, so it may be that we sometimes can't understand what you're saying because of language barrier. Mar 3 '18 at 16:25
• Also, for the former challenge, I left a comment "Then how would you score different programs?" -- and yes, that's a problem. Mar 3 '18 at 16:27
• True I'm not native English speaker, but even when saying mother language I don't say that right(only enough to understand). Your "how score" only happen when two are near, which now seem don't happen
– l4m2
Mar 3 '18 at 16:33
• @l4m2 it still is counted as unclear even if the question at hand doesn't matter right now, as a challenge should be continuable without the OP. Mar 3 '18 at 16:36
• @dzaima What do you want to say? / But fastest-code challenges need the OP to judge programs. Mar 3 '18 at 16:37
• @user202729 fastest-code requires the OP for all submissions all the time for a specific thing, which is not the case here, and it isn't specified how the comparing by the OP will happen Mar 3 '18 at 16:44
• @dzaima Still, it's possible to require answers to prove their (relative) score. Mar 3 '18 at 16:48
• @user202729 which (unless I'm blind) the challenge isn't currently asking for. Mar 3 '18 at 16:50
• The former problem Node 5 has some bug(defining float with k-bit precise and left are exp bits, and 1/eps only matter with k). When problem bugged what is that?
– l4m2
Mar 3 '18 at 16:52
• @l4m2 What is "problem Node 5"? Mar 3 '18 at 17:03
• Problem 1 Node 5
– l4m2
Mar 3 '18 at 23:52

## 2 Answers

"Unclear" can mean any of the following:

• I don't understand the specification well enough to judge whether a program meets it or not (so by implication, I cannot answer it), and that's not just because I don't know enough of some obscure branch of mathematics which is an integral part of the question.
• I think I do understand the specification, and it's self-contradictory.
• I can see incompatible interpretations of the specification (i.e. it's ambiguous).
• I'm convinced that the specification is unambiguous, someone else is also convinced that it's unambiguous, but our interpretations disagree.
• I'm convinced that the specification is unambiguous, but I'm also pretty sure that what you're asking for isn't what you intended to ask for. (E.g. you ask me to output the 17th integer in a sequence, and the obvious answer is to hard-code it rather than to calculate it).

In your specific case, the question had several problems that probably should make it closed. Users are more likely only unsure of whether it is salvageable, and if not, which one is the real unsalvageable reason to close it. And they didn't find the answer from your replies in the comments.

Now I get the idea that this model is stronger than a Turing machine, as it can compute a "limit" in infinite time. If this is exactly your intention and the reason you find it interesting, better answer this to the question "why ask this". And explain this exactly, instead of linking to uncommented source code on external sites. And other users may help cleaning up the other parts. (But if you intend to revive this question, better use the sandbox.) But it has too many other problems that this looks like a coincidence.

As an example, x86 is an instruction set. There is no such thing as 128-bit x86, because nobody has invented the opcodes for the new instructions yet, even though your idea about what it is like might not be too far. Depending on what you intended to mean exactly, this might be easy to fix, or not.

Some obvious options are to add a full specification of x86 with variable register size, or simply change n to an input and not run into the problem of defining number types. The former needs a lot of work and may not make it on-topic. The later changes the question significantly and would make people rethink whether it is a duplicate. Nobody is likely going to do these unless they are sure their interpretation is right.

In your situation, you shouldn't consider it a minor, irrelevant, technical problem that could be ignored. It is a probably minor problem that makes the question in its current form should absolutely be closed with no excuse. It's only less often so because editing it could be easy, if someone else knows your intention. But if there are 5 different parts in the question that it is 80% "obviously" sure what it means each, another single interpretation that is only 35% sure, completely changing the problem and making the 5 parts irrelevant is more likely, even if it assumes you have made more mistakes similar to the x86 problem while an interpretation without those is possible.

Worse, that you said unclear not seeming to mean "don't know what the question mean", so I would easily consider my interpretation overthinking, and you agree the user who edited it know what it means and did it right.

So, it is just unclear. Not answering the questions only leaves it in an unclear state, however big or small the problems are. Better firstly make it clear what's the most important part and why a rule must be there than insisting on everything. Acknowledging which users interpreted it correctly or wrongly could help further. Also note that even if everybody knows your intention, it could still be difficult to get the rules exactly right for some types of challenges.