20
$\begingroup$

I searched for questions on perfect numbers. Most of the questions are abbreviated, but it looks like Tharsis found a workaround:

enter image description here

What's going on here?

$\endgroup$
  • $\begingroup$ That's odd. It chopped off at the usual spot for me. Edit With the search string "perfect number" (with quotes) it indeed shows up in its entirety. $\endgroup$ – user714630 Sep 16 '13 at 21:28
  • $\begingroup$ Actually every entry is a complete post when I use that search string. $\endgroup$ – user714630 Sep 16 '13 at 21:36
  • $\begingroup$ @KarlKronenfeld strange. BTW the image is also a link, to the search with quotes... $\endgroup$ – draks ... Sep 16 '13 at 21:45
  • $\begingroup$ Ah, I didn't notice that, thanks. Playing around with the strings pure, "pure", and "pure subgroup" I noticed that the first two had drastically different search results but that the posts were chopped off. The last search displayed complete posts. $\endgroup$ – user714630 Sep 16 '13 at 21:51
  • $\begingroup$ @karl maybe it's feature not a bug... $\endgroup$ – draks ... Sep 16 '13 at 21:53
  • $\begingroup$ Searching for "odd perfect number" returns two posts with the same problem. $\endgroup$ – Martin Sleziak Sep 17 '13 at 6:54
  • 1
    $\begingroup$ This seems rather similar to the following bug report on meta.SO: Whoa! The search results just gave me a book! $\endgroup$ – Martin Sleziak Sep 17 '13 at 7:38
  • 3
    $\begingroup$ As you can see from the bold highlighting, the post contains your search term multiple times, and mostly formulae in between. It seems like the search view displays all occurrences of the search term in a post plus some surrounding text (which seems to more than just double when matches are closely together), and that the presence of formulae seems to somehow extend this surrounding. Maybe a specific character breaks the "am I still close to a match?" function $\endgroup$ – Tobias Kienzler Sep 17 '13 at 11:59
  • 2
    $\begingroup$ @TobiasKienzler if you look at the search linked in Martin's comment (direct link), you'll only find one occurence of "code review" at the beginning of the post... $\endgroup$ – draks ... Sep 17 '13 at 20:12
  • 1
    $\begingroup$ @draks Hm, indeed. I noticed that all the bloated results are link texts, so maybe one of the characters involved there is responsible. In the "perfect number" case, there's a link directly following one mention of the term, let $N$ be an odd perfect number. It is [not difficult](http://math.stackexchange.com/q/339894/53846) to show. So I'd suspect markdown links are involved in the reason for this. We could test this by posting an answer here twice with one made-up unique keyword, once followed by a link and once without $\endgroup$ – Tobias Kienzler Sep 18 '13 at 8:09
  • $\begingroup$ @TobiasKienzler ok, done, but the test doesn't work well: (1) searching without quotes doesn't find anything at all, see here; (2) searching without doesn't make a difference, see here...:-( $\endgroup$ – draks ... Sep 18 '13 at 20:35
  • $\begingroup$ Hmm, it's not the links, see here: it doesn't make a difference with or without links in the search view... $\endgroup$ – draks ... Sep 18 '13 at 20:42
  • $\begingroup$ @draks... What I find surprising is that only two of the four answers show up in the results. The other posts contain "perfect number" multiple times as well. $\endgroup$ – Lord_Farin Sep 19 '13 at 6:36
  • $\begingroup$ This is really weird. I hope one of the devs does have a look into this at some point... $\endgroup$ – Tobias Kienzler Sep 19 '13 at 7:11
  • $\begingroup$ feel free to add your test answer... $\endgroup$ – draks ... Oct 19 '13 at 21:24
1
$\begingroup$

Test answer with links

let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc. Curabitur eget dictum massa. Sed ac nunc eu elit fermentum condimentum nec sed dui. Duis vel semper enim, ut molestie leo. Donec viverra velit enim. Nulla congue fermentum aliquam. Sed a arcu elit. Nulla sollicitudin eget urna a ultricies. Donec varius blandit augue, a mattis ipsum bibendum sed. Nullam auctor luctus risus sed eleifend. Proin pulvinar nisi pulvinar, tristique velit vitae, adipiscing leo. Nunc nisl orci, pulvinar nec tellus a, cursus consequat nunc.

Nullam semper velit quis massa fringilla rutrum. Nam et metus nec lorem mollis bibendum. Fusce sem justo, congue pharetra ultrices sit amet, pulvinar id leo. Donec posuere in lorem at rhoncus. Vestibulum mollis vestibulum posuere. Nullam dapibus euismod nulla, sed mattis enim lobortis commodo. Etiam luctus tellus ut tellus cursus, sed sollicitudin augue vehicula. Nunc condimentum, nibh eget consectetur ornare, elit metus tempus sapien, non fringilla sem nulla ut purus. Nulla id tellus porttitor, tristique sem sit amet, imperdiet orci. Morbi vehicula lobortis velit in facilisis. Integer a sollicitudin eros.

let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc. Curabitur eget dictum massa. Sed ac nunc eu elit fermentum condimentum nec sed dui. Duis vel semper enim, ut molestie leo. Donec viverra velit enim. Nulla congue fermentum aliquam. Sed a arcu elit. Nulla sollicitudin eget urna a ultricies. Donec varius blandit augue, a mattis ipsum bibendum sed. Nullam auctor luctus risus sed eleifend. Proin pulvinar nisi pulvinar, tristique velit vitae, adipiscing leo. Nunc nisl orci, pulvinar nec tellus a, cursus consequat nunc.

Nullam semper velit quis massa fringilla rutrum. Nam et metus nec lorem mollis bibendum. Fusce sem justo, congue pharetra ultrices sit amet, pulvinar id leo. Donec posuere in lorem at rhoncus. Vestibulum mollis vestibulum posuere. Nullam dapibus euismod nulla, sed mattis enim lobortis commodo. Etiam luctus tellus ut tellus cursus, sed sollicitudin augue vehicula. Nunc condimentum, nibh eget consectetur ornare, elit metus tempus sapien, non fringilla sem nulla ut purus. Nulla id tellus porttitor, tristique sem sit amet, imperdiet orci. Morbi vehicula lobortis velit in facilisis. Integer a sollicitudin eros.

let $N$ be an odd perfect number. It is not difficult to show.

$\endgroup$
0
$\begingroup$

Let $\omega(n)$ denote the number of distinct prime factors of a positive integer $n$, and let $N$ be an odd perfect number. It is not difficult to show that $\omega(N)\ge3$. In fact, Nocco already proved this in 1863. Showing that $\omega(N)\neq3$ requires a little more effort, but the proof is still fairly short. I will present some version of the proof below; the original one is due to Peirce in 1830 (which, interestingly, seems to have been published before Nocco's proof).

The two key ideas here are the geometric series and the factor chain method. Let $N=\displaystyle\prod_{i=1}^k p_i^{e_i}$ be the prime factorization of $N$. Then it follows that

$$\sigma(N)=\prod_{i=1}^k \sum_{j=0}^{e_i} p_i^j=\prod_{i=1}^k \frac{p_i^{e_i+1}-1}{p_i-1}.$$

This formula is useful, because each factor is necessarily greater than one when $e_i$ is positive, which is the case for all prime factors $p_i$. Hence $\sigma(N)\ge\sigma\left(\displaystyle\prod_{i\in{I}} p_i^{e_i}\right)$, where $I\subseteq\{1,2,\ldots,k\}$.

The second idea is perhaps even more fundamental, because we have yet to use the property of perfectness. Since $\sigma(N)=2N$, every divisor of $\sigma(N)$ must also divide $2N$; the other way around is not as interesting. In particular, every factor $\frac{p_i^{e_i+1}-1}{p_i-1}$ must divide $2N$. This result is really powerful; for example, observing that $2^1||2N$ yields Euler's well-known form of an odd perfect number after some algebra.

With these tools in our toolbox, the proof is almost immediate. Let $p^\alpha$, $q^\beta$, and $r^\gamma$ be three odd prime powers with $p<q<r$ and assume that $N=p^\alpha q^\beta r^\gamma$ is an odd perfect number. By the same token as above, we see that

$$\begin{align*}\sigma(N)=\sigma(p^\alpha q^\beta r^\gamma) &= \frac{p^{\alpha+1}-1}{p-1} \cdot \frac{q^{\beta+1}-1}{q-1} \cdot \frac{r^{\gamma+1}-1}{r-1} < \frac{p^\alpha p}{p-1} \cdot \frac{q^\beta q}{q-1} \cdot \frac{r^\gamma r}{r-1}\\\implies& \frac{\sigma(N)}{N} = \frac{\sigma(p^\alpha q^\beta r^\gamma)}{p^\alpha q^\beta r^\gamma} < \frac{p}{p-1} \cdot \frac{q}{q-1} \cdot \frac{r}{r-1}.\end{align*}$$

Suppose that $p=5$. Then $\frac{\sigma(N)}{N}<\frac{5}{4}\cdot\frac{7}{6}\cdot\frac{11}{10}=\frac{77}{48}<2$, which is impossible. Hence $p=3$.

Suppose that $q=7$. Then $\frac{\sigma(N)}{N}<\frac{3}{2}\cdot\frac{7}{6}\cdot\frac{11}{10}=\frac{77}{40}<2$, which is impossible. Hence $q=5$.

Suppose that $r=17$. Then $\frac{\sigma(N)}{N}<\frac{3}{2}\cdot\frac{5}{4}\cdot\frac{17}{16}=\frac{255}{128}<2$, which is impossible. Hence $r \le 13$.

In consequence, we have three possible cases to consider: $r=7$, $r=11$, and $r=13$.

In the first case, we have $N=3^\alpha 5^\beta 7^\gamma$. By the factor chain method, $3^1||N$ implies $4|2N$ and $7^1||N$ implies $8|2N$, neither of which is possible. Hence $\alpha\ge2$ and $\gamma\ge2$. It follows that

$$\frac{\sigma(N)}{N} \ge \frac{\sigma(3^2\cdot5\cdot7^2)}{3^2\cdot5\cdot7^2} = \left(1 + \frac{1}{3} + \frac{1}{3^2}\right)\left(1 + \frac{1}{5}\right)\left(1 + \frac{1}{7} + \frac{1}{7^2}\right)=\frac{494}{245}>2,$$

which is impossible, but we already knew that $105=3\cdot5\cdot7$ cannot divide an odd perfect number.

In the second case, we have $N=3^\alpha 5^\beta 11^\gamma$. Suppose that $\beta=1$. It follows that

$$\frac{\sigma(N)}{N} = \frac{\sigma(3^\alpha\cdot5\cdot11^\gamma)}{3^\alpha\cdot5\cdot11^\gamma} < \frac{3}{2} \cdot \frac{6}{5} \cdot \frac{11}{10} = \frac{99}{50} < 2,$$

which is impossible (could we know that beforehand?). Hence $\beta\ge2$. By the factor chain method, $\alpha\ge4$ and $\gamma\ge4$. It follows that

$$\frac{\sigma(N)}{N} \ge \frac{\sigma(3^4\cdot5^2\cdot11^4)}{3^4\cdot5^2\cdot11^4} = \left(1 + \frac{1}{3} + \frac{1}{3^2} + \frac{1}{3^3} + \frac{1}{3^4}\right)\left(1 + \frac{1}{5} + \frac{1}{5^2}\right)\left(1 + \frac{1}{11} + \frac{1}{11^2} + \frac{1}{11^3} + \frac{1}{11^4}\right)=\frac{99851}{49005}>2,$$

which is impossible, but we already knew that $825=3\cdot5^2\cdot11$ cannot divide an odd perfect number.

In the third case, we have $N=3^\alpha 5^\beta 13^\gamma$. $\beta=1$ implies deficiency, so $\beta\ge2$. Now, $\alpha=2$ also implies deficiency, so $\alpha\ge4$. By the factor chain method, we must have $\gamma\ge2$, which this time implies abundance. Hence the assumption leads to a contradiction; furthermore, $8775=3^3\cdot5^2\cdot13$ cannot divide an odd perfect number. We are finally done.

Unfortunately, proving that $\omega(N)\neq4$ is not this easy by hand. We can similarly show that $p=3$, but then we already have two possible choices for $q$. Of course, one could just write a program and loop through all the possibilities, but Sylvester did it successfully in 1888, unlikely by the aid of computers. Hence I wonder if this algorithm can be significantly improved, or if there are other more efficient techniques that achieve the same goal.

$\endgroup$
0
$\begingroup$

Let $\omega(n)$ denote the number of distinct prime factors of a positive integer $n$, and let $N$ be an odd perfect number. It is not difficult to show that $\omega(N)\ge3$. In fact, Nocco already proved this in 1863. Showing that $\omega(N)\neq3$ requires a little more effort, but the proof is still fairly short. I will present some version of the proof below; the original one is due to Peirce in 1830 (which, interestingly, seems to have been published before Nocco's proof).

The two key ideas here are the geometric series and the factor chain method. Let $N=\displaystyle\prod_{i=1}^k p_i^{e_i}$ be the prime factorization of $N$. Then it follows that

$$\sigma(N)=\prod_{i=1}^k \sum_{j=0}^{e_i} p_i^j=\prod_{i=1}^k \frac{p_i^{e_i+1}-1}{p_i-1}.$$

This formula is useful, because each factor is necessarily greater than one when $e_i$ is positive, which is the case for all prime factors $p_i$. Hence $\sigma(N)\ge\sigma\left(\displaystyle\prod_{i\in{I}} p_i^{e_i}\right)$, where $I\subseteq\{1,2,\ldots,k\}$.

The second idea is perhaps even more fundamental, because we have yet to use the property of perfectness. Since $\sigma(N)=2N$, every divisor of $\sigma(N)$ must also divide $2N$; the other way around is not as interesting. In particular, every factor $\frac{p_i^{e_i+1}-1}{p_i-1}$ must divide $2N$. This result is really powerful; for example, observing that $2^1||2N$ yields Euler's well-known form of an odd perfect number after some algebra.

With these tools in our toolbox, the proof is almost immediate. Let $p^\alpha$, $q^\beta$, and $r^\gamma$ be three odd prime powers with $p<q<r$ and assume that $N=p^\alpha q^\beta r^\gamma$ is an odd perfect number. By the same token as above, we see that

$$\begin{align*}\sigma(N)=\sigma(p^\alpha q^\beta r^\gamma) &= \frac{p^{\alpha+1}-1}{p-1} \cdot \frac{q^{\beta+1}-1}{q-1} \cdot \frac{r^{\gamma+1}-1}{r-1} < \frac{p^\alpha p}{p-1} \cdot \frac{q^\beta q}{q-1} \cdot \frac{r^\gamma r}{r-1}\\\implies& \frac{\sigma(N)}{N} = \frac{\sigma(p^\alpha q^\beta r^\gamma)}{p^\alpha q^\beta r^\gamma} < \frac{p}{p-1} \cdot \frac{q}{q-1} \cdot \frac{r}{r-1}.\end{align*}$$

Suppose that $p=5$. Then $\frac{\sigma(N)}{N}<\frac{5}{4}\cdot\frac{7}{6}\cdot\frac{11}{10}=\frac{77}{48}<2$, which is impossible. Hence $p=3$.

Suppose that $q=7$. Then $\frac{\sigma(N)}{N}<\frac{3}{2}\cdot\frac{7}{6}\cdot\frac{11}{10}=\frac{77}{40}<2$, which is impossible. Hence $q=5$.

Suppose that $r=17$. Then $\frac{\sigma(N)}{N}<\frac{3}{2}\cdot\frac{5}{4}\cdot\frac{17}{16}=\frac{255}{128}<2$, which is impossible. Hence $r \le 13$.

In consequence, we have three possible cases to consider: $r=7$, $r=11$, and $r=13$.

In the first case, we have $N=3^\alpha 5^\beta 7^\gamma$. By the factor chain method, $3^1||N$ implies $4|2N$ and $7^1||N$ implies $8|2N$, neither of which is possible. Hence $\alpha\ge2$ and $\gamma\ge2$. It follows that

$$\frac{\sigma(N)}{N} \ge \frac{\sigma(3^2\cdot5\cdot7^2)}{3^2\cdot5\cdot7^2} = \left(1 + \frac{1}{3} + \frac{1}{3^2}\right)\left(1 + \frac{1}{5}\right)\left(1 + \frac{1}{7} + \frac{1}{7^2}\right)=\frac{494}{245}>2,$$

which is impossible, but we already knew that $105=3\cdot5\cdot7$ cannot divide an odd perfect number.

In the second case, we have $N=3^\alpha 5^\beta 11^\gamma$. Suppose that $\beta=1$. It follows that

$$\frac{\sigma(N)}{N} = \frac{\sigma(3^\alpha\cdot5\cdot11^\gamma)}{3^\alpha\cdot5\cdot11^\gamma} < \frac{3}{2} \cdot \frac{6}{5} \cdot \frac{11}{10} = \frac{99}{50} < 2,$$

which is impossible (could we know that beforehand?). Hence $\beta\ge2$. By the factor chain method, $\alpha\ge4$ and $\gamma\ge4$. It follows that

$$\frac{\sigma(N)}{N} \ge \frac{\sigma(3^4\cdot5^2\cdot11^4)}{3^4\cdot5^2\cdot11^4} = \left(1 + \frac{1}{3} + \frac{1}{3^2} + \frac{1}{3^3} + \frac{1}{3^4}\right)\left(1 + \frac{1}{5} + \frac{1}{5^2}\right)\left(1 + \frac{1}{11} + \frac{1}{11^2} + \frac{1}{11^3} + \frac{1}{11^4}\right)=\frac{99851}{49005}>2,$$

which is impossible, but we already knew that $825=3\cdot5^2\cdot11$ cannot divide an odd perfect number.

In the third case, we have $N=3^\alpha 5^\beta 13^\gamma$. $\beta=1$ implies deficiency, so $\beta\ge2$. Now, $\alpha=2$ also implies deficiency, so $\alpha\ge4$. By the factor chain method, we must have $\gamma\ge2$, which this time implies abundance. Hence the assumption leads to a contradiction; furthermore, $8775=3^3\cdot5^2\cdot13$ cannot divide an odd perfect number. We are finally done.

Unfortunately, proving that $\omega(N)\neq4$ is not this easy by hand. We can similarly show that $p=3$, but then we already have two possible choices for $q$. Of course, one could just write a program and loop through all the possibilities, but Sylvester did it successfully in 1888, unlikely by the aid of computers. Hence I wonder if this algorithm can be significantly improved, or if there are other more efficient techniques that achieve the same goal.

$\endgroup$
0
$\begingroup$

Test answer without links

EDIT now with quotes

let $N$ be an odd "perfect number". It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc. Curabitur eget dictum massa. Sed ac nunc eu elit fermentum condimentum nec sed dui. Duis vel semper enim, ut molestie leo. Donec viverra velit enim. Nulla congue fermentum aliquam. Sed a arcu elit. Nulla sollicitudin eget urna a ultricies. Donec varius blandit augue, a mattis ipsum bibendum sed. Nullam auctor luctus risus sed eleifend. Proin pulvinar nisi pulvinar, tristique velit vitae, adipiscing leo. Nunc nisl orci, pulvinar nec tellus a, cursus consequat nunc.

Nullam semper velit quis massa fringilla rutrum. Nam et metus nec lorem mollis bibendum. Fusce sem justo, congue pharetra ultrices sit amet, pulvinar id leo. Donec posuere in lorem at rhoncus. Vestibulum mollis vestibulum posuere. Nullam dapibus euismod nulla, sed mattis enim lobortis commodo. Etiam luctus tellus ut tellus cursus, sed sollicitudin augue vehicula. Nunc condimentum, nibh eget consectetur ornare, elit metus tempus sapien, non fringilla sem nulla ut purus. Nulla id tellus porttitor, tristique sem sit amet, imperdiet orci. Morbi vehicula lobortis velit in facilisis. Integer a sollicitudin eros.

let $N$ be an odd "perfect number". It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc. Curabitur eget dictum massa. Sed ac nunc eu elit fermentum condimentum nec sed dui. Duis vel semper enim, ut molestie leo. Donec viverra velit enim. Nulla congue fermentum aliquam. Sed a arcu elit. Nulla sollicitudin eget urna a ultricies. Donec varius blandit augue, a mattis ipsum bibendum sed. Nullam auctor luctus risus sed eleifend. Proin pulvinar nisi pulvinar, tristique velit vitae, adipiscing leo. Nunc nisl orci, pulvinar nec tellus a, cursus consequat nunc.

Nullam semper velit quis massa fringilla rutrum. Nam et metus nec lorem mollis bibendum. Fusce sem justo, congue pharetra ultrices sit amet, pulvinar id leo. Donec posuere in lorem at rhoncus. Vestibulum mollis vestibulum posuere. Nullam dapibus euismod nulla, sed mattis enim lobortis commodo. Etiam luctus tellus ut tellus cursus, sed sollicitudin augue vehicula. Nunc condimentum, nibh eget consectetur ornare, elit metus tempus sapien, non fringilla sem nulla ut purus. Nulla id tellus porttitor, tristique sem sit amet, imperdiet orci. Morbi vehicula lobortis velit in facilisis. Integer a sollicitudin eros.

let $N$ be an odd "perfect number". It is not difficult to show.

$\endgroup$
0
$\begingroup$

Test answer with links

let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.let $N$ be an odd perfect number. It is not difficult to show.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut sed pharetra erat, in lacinia nunc.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .