# What is the problem with this post? [closed]

I asked this question: How to find the fraction of integers with the smallest denominator matching an interval?

Maybe you looked over something?

What "context or other details" are missing?

What is going on here?

• Did you read this: "This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level." – user99914 Jul 19 '18 at 0:32
• @JohnMa Yes, I read. I can't remember, how I tried to solve it. The simple and obvious iterative search has fallen out from the first moment, it is not enough effective. My other ideas failed because I could not make sure if they find really the smallest $k$. Do you think it would be a real improvement if I would write these there? I think it would be more like noise. – peterh Jul 19 '18 at 0:35
• I have voted to reopen. This is a serious question in Number Theory, not some textbook exercise in high school algebra, and I think an optimal answer requires some knowledge of continued fractions. My only qualm is that it's quite likely that a question very like it has been asked and answered already, so it might be worth searching for a duplicate. – Gerry Myerson Jul 19 '18 at 0:45
• it's now deleted. – user99914 Jul 19 '18 at 0:48
• If you're still interested in answers, peterh, type $$\rm decimal\ to\ fraction\ site:math.stackexchange.com$$ into the internet and you should find many helpful discussions of previous versions of your question. – Gerry Myerson Jul 19 '18 at 1:03
• It is now undeleted. – José Carlos Santos Jul 19 '18 at 8:31
• Probably lack of displayed effort. However in my opinion, lack of effort is not always a problem. See also: What have you tried? ... – user202729 Jul 19 '18 at 9:41
• @GerryMyerson Thank you very much! I've checked for dupe, but didn't find anything. However, it is not easy to search for dupes of MathSE posts. – peterh Jul 19 '18 at 9:51
• @user202729 I think I could write these into the post, but it would work essentially as a noise. This post originated as a codegolf problem (referenced in the first comment of the post), I posted its mathematical part here, because I considered also its mathematical side particularly interesting. It was a l'art pour l'art "what beautiful problem is it" post, not a "do my homework instead me" post! – peterh Jul 19 '18 at 10:32
• @GerryMyerson: I honestly don't see why this is a number theory question. For an interval $[c,d[$ one just try all natural number $\le \frac{1}{d-c}$. If there is anything really interesting, please do add to the post. – user99914 Jul 19 '18 at 10:33
• @JohnMa This is like saying, "to find the least common multiple of $a$ and $b$, just try all the natural numbers $\lt a\cdot b$... while it is literally true, it is not a good answer. – peterh Jul 19 '18 at 10:37
• math.stackexchange.com/q/1037286/29335 is very nearly a duplicate, but I see it deals with a closed interval instead. There are also several dealing with looking for the best rational between two other rationals. – rschwieb Jul 19 '18 at 10:38
• @JohnMa Btw, if you think it is not a number theory question, then the correct step is to re-tag it (or suggest its re-tagging in an edit), and not a down-close-del vote. – peterh Jul 19 '18 at 10:39
• @peterh Why is it not a good answer? – user99914 Jul 19 '18 at 10:40
• @peterh I don't see any signal-noise ratio problems in the other question. I would evaluate them as being approximately equal in content and quality. – rschwieb Jul 19 '18 at 13:21

1. I think that the question is not of great quality. It is essentially a "Problem Statement Question"—you have presented a problem, but you have not given anyone any reason to believe that it is an interesting problem, you have not explained where the problem comes from, and you have not described the kinds of tools that you think will be necessary to solve the problem. You haven't even explained why you think that there is a faster algorithm.

Remember that the goal of MSE is to collect questions about mathematics together with answers. Bare problem statements, such as those that you might find in a textbook, are problems, not questions. The goal of MSE is not to create a solutions manual to "all the problems in math". You need to make it clear somehow that your question is about more than just a problem that you don't know how to solve.

2. That being said, it seems clear that such context could be provided for this problem, and you have (in the last couple of hours) added that context, vastly improving the question. I'm still not sure I understand why anyone should care about the question (i.e. it still feels very unmotivated to me, despite Gerry Myerson's protestations in the comments above), but I am much more satisfied with the current form of the question than I was with the bare problem statement that I voted to close in June.

3. However, it seems that rschwieb has found a good target for duplication: Finding the simplest rational in a closed interval (I would be really curious to know how you found that, rschwieb—my 15 minutes with Google and approach0 didn't turn up anything). My opinion is that, at this moment in time, the appropriate action to take is to mark the newer question as a duplicate of the older question. Since the older question has an answer that is virtually identical to the newer answer, this seems like a reasonable course of action.

Note that the older question also has many of the flaws described above, but the last several times that I have asked to get older questions marked as duplicates of newer questions, I've gotten pushback. It seems that the consensus is that newer questions should be marked as duplicates of older questions.

• I cut and paste the user's question title into approach0.xyz and that was the first hit for me: here (Of course, this may change over time due to data changes.) It took all of 5 seconds. – rschwieb Jul 19 '18 at 13:47
• @rschwieb Wow... I was overthinking things way too much. I was looking for TeX'd bits. Thanks! – Xander Henderson Jul 19 '18 at 14:40
• Often the convention of preferring older to newer with duplicates can be overturned if the newer question is significantly better written, particularly when they're closely worded enough for an easy merge. – Alexander Gruber Jul 19 '18 at 15:36