Hacker News new | past | comments | ask | show | jobs | submit login
The Shortest Papers Ever Published (2016) (paperpile.com)
382 points by diodorus on Nov 20, 2017 | hide | past | web | favorite | 93 comments

Although these short papers are fun, I think the most interesting papers are those that are both important and concise. One of my favorite examples is Josh Nash's proof of the existence of Nash Equilibria (which is a foundational concept in game theory and arguably won Nash a Nobel):


It's less than a full page in the original journal.

His Ph.D. thesis is also relatively short (just 28 pages):


There's a related, typeset version from the Annals of Mathematics five months later, on page 9 of:


While we're on the subject of brevity, page 23 has the shortest recommendation letter I've ever seen.

Is this some sort of draft? I've seen manuscripts from the 30s that are much more legible.

My supervisor's thesis from the 70's is in about as good as shape as the one linked. Must depend on when it was scanned.

Agreed. In a few days I'm supposed to give a lecture about how to write great scientific papers. I wasn't aware of this paper (I'm aphysicist, not a mathematician), but nevertheless I'm going to discuss it with the class, its conciseness is exemplary! Thanks a lot!

Can you post your presentation here?? I would be interested to see it.

I would be happy to put it on slideshare, but it's going to be written in italian…

Be sure to check this video, as it is very good: https://www.youtube.com/watch?v=g3dkRsTqdDA&index=38&list=FL...

The discovery of DNA was also a page and yielded a Nobel prize


The discovery of the structure of DNA*. DNA had been known as one of the components of cells for a long time, and the Avery, Macleod, and McCarty paper from 1944[0] showed that DNA was the substance that carried information, rather than protein.

[0] http://jem.rupress.org/content/79/2/137

Nature appears to do some silly referrer blocking.

It's the first link under "archive" on https://www.nature.com/nature/dna50/archive.html

That wasn't working for me either.

Here's an IPFS link: https://www.eternum.io/ipfs/QmaCQy4GdRTh6jHf2eDJVoWzgHzTbwVj...

I always loved this line:

"It has not escaped our notice that the specific pairing we have postulated immediately suggests a possible copying mechanism for the genetic material."

(Your link appears broken, btw)

I love this example which shows how a "novel" publication can be that simple and beautiful.

One of the shortest papers ever, or maybe the longest:


If you want to see what he did there, read again from the start of this sentence.

This guy solved two open problems in mathematics because he mistook them for a homework assignment:


"A year later, when I began to worry about a thesis topic, Neyman just shrugged and told me to wrap the two problems in a binder and he would accept them as my thesis."

Each paper was seven pages, which would make a fourteen page thesis.

Very, cool! In fact, a paper which solved one of these problems [1] is precisely 7 pages. Additionally, it contains the footnote: "The main results of this paper were obtained by the authors independently of each other using entirely different methods."

[1] https://projecteuclid.org/euclid.aoms/1177729695

What's ironic in itself is that the mentioned paper ("A comprehensive overview of chemical-free consumer products", published by Wiley; i.e. an empty page) costs $38 to download as PDF: http://onlinelibrary.wiley.com/wol1/doi/10.1002/ciuz.2016007...

This makes it the most expensive piece of literature (cost per word) that I know of.

Those journals got to recoup their expenses and hard work. /s

I suppose they just set a default price for all papers, and manually adjust others or something like that? I’m sure it’s not as silly as it seems right

I'm sure they could price by number of pages.

Since it's zero words long, it would be the most expensive piece of literature no matter what price they set on it!

Unfortunately, most scientific and scholarly literature is among the most expensive pieces of literature.

“Scientific.” “Scholarly.”

Given its, uh, content, this particular paper can really be classified as anything.


i.e. I wasn't talking about this specific paper, but most of all the other papers.

Indeed you were not. However, given that you generalized about the categories within which this paper presumably lies, I felt it amusing to point out that, using this paper as a start, one could generalize about any particular categories of papers.

Oh, that might just be the result of my not being a native speaker. In my language, the term "scientific" refers to articles in both the Humanities and Social Sciences as well as what Science refers to in English. Hence, I usually feel the need to specify all of those. In other words: it had nothing to do with the mentioned article being part of those domains :)

How about the most silent lecture?

"Cole’s lecture was different. He did not speak a single word. He simply went to the board, and began to calculate. On one side of the board, he calculated 267 – 1 = 147,573,952,589,676,412,927 by hand. Then he went to the other side of the board and worked out the product of 193,707,721 and 761,838,257,287, the factors of 147,573,952,589,676,412,927. After spending the silent hour working out the calculations, Cole simply turned around and went back to his seat, completely silent! The audience erupted into a standing ovation."


> 267 – 1 = 147,573,952,589,676,412,927 by hand

I think I spotted an error there.

It should be 2^67 - 1.

That was the error.

I erupted into a standing ovation before Cole even went to the other side of the board.

> 2047 is a composite number with factors 23 and 87.

This is another error (though this one is actually in the article). The factors are 23 and 89.

A bit tangential but amusing: my PhD dissertation was 51 pages, all math. One of my committee members approved of the work but wanted the dissertation to be bulked up with a program listing- thankfully my advisor firmly rejected this suggestion :-)

I fail to understand how the content in the Soifer paper (triangles) answers its title question. So yes, some explanation would have been necessary in my opinion.

That's the thing, most of these are actually pretty bad papers. Their editors were right, it would have been improved with at least a couple of sentences of background info or explanation. For example, why is that problem even an interesting one to ask? Should we expect the answer to the titular question to be no? Is there anything particularly novel or interesting about the existence proof they've provided that could maybe be applied to other problems? And so on.

I disagree: the paper is clearly not trying to answer it's titular question, in fact it is justifying why that question is interesting to ask. In particular, it demonstrates that the answer to the obvious other question to ask (with n+2 instead of n+1) is yes and clearly if you altered the question the other direction (with n instead of n+1), then the answer is no. This immediately makes me find the question compelling.

Papers motivating problems are, frankly, quite important and should be done more often.

Well, that way around it makes sense. But what you wrote here could have been written in the paper, so people like me would have understood it as well.

Yeah, it's definitely arguable. They clearly were trying to be cute with the paper. Though note that the published version of the paper was slightly longer (but without really enhancing its clarity).

My Favorite is Dan Janzen's paper "Yes?" that was published in Biotropica in 1978. Here it is in full:



    Acknowledgments: This study was supported by NSF DEB77-04889, 
    and grew out of discussion with the Society for Historical Orations on Theory.

    Daniel H. Janzen
    Department of Biology,
    University of Pennsylvania,
    Philadelphia, Pennsylvania 19104, USA

This is taking Betteridge's Law of Headlines a bit far...

I just had to look this one up though, so for posterity: https://doi.org/10.2307/2387687.

I want to know what the peer reviews looked like.



There's an excellent Numberphile on this topic, most of the papers in the post are mentioned in the video: https://www.youtube.com/watch?v=QvvkJT8myeI

In my opinion, it is clear the author simply copied that video, because not only is it the exact same set of papers, if memory serves (as I watched it only last week), it is in the exact same order, which is particularly determinative. Citation would be appropriate at the very least.

A citation would absolutely be appropriate. However, you have it the wrong way around.

This blog post was published Jun 17, 2016.

The Numberphile video was published Dec 7, 2016.

I stand corrected then. And disappointed... I like the Numberphile videos.

Reminds me of http://www.tinytocs.org, computer science papers whose bodies must be 140 characters or less.

twitter henceforth

On a related note, I'm fond of this Mathoverflow answer, "Which math paper maximizes the ratio (importance)/(length)?"[1] Highlights include:

1. A one-sentence proof published in American Mathematical Monthly that costs $19 to download on JSTOR,

2. The 8-page paper that introduced ζ (zeta) notation, two proofs of the ζ(s) L-function, several new methods in analytic and number theory, and (most famously) the Riemann Hypothesis,

3. Lebesque's paper introducing modern measure theory,

4. Elkies' paper proving the (titular) existence of infinitely many supersingular primes for every elliptic curve defined on the rationals.

It seems it was a bit easier to publish short, dense material in the 20th century :)


1. https://mathoverflow.net/questions/7330/which-math-paper-max...

This reminds me of 'El Dinosaurio,' the shortest Spanish narrative:

"Cuando despertó, el dinosaurio todavía estaba allí."


"When he woke up, the dinosaur was still there."

I like the idea of flash non-fiction, but I never seem able to find much about it. Not one-liners, rather in the 3-5 paragraph range, which might be understood as being an article (ex. newspaper/magazine), only articles are usually longer.

I'd love to find an example of a media outlet that does such a thing, albeit with mid/high-brow content.

365 tomorrows might be the right length for you: https://365tomorrows.com/ (only science fiction)

http://sixsentences.blogspot.ch/ is pretty good sometimes.

I once had a directory with huge number of fiction books, and I've sorted it by size asc. Most of the first few dozens of entries were amusing.

Which in turn reminds me of Hemingway's shortest short story:

'For sale: baby shoes, never worn.'

There's no evidence that it was Hemingway who wrote it.


Gareth Evans' 1978 1-pager "Can There Be Vague Objects?" is notable both for its brevity, and for its significance.


One of my favorite papers is a pretty short but famous one: Is Justified True Belief Knowledge?


How bizarre is it that I've read this paper before. I think it's provocative, but, this not being my field, I don't know what effect it had.

It's one of the most influential papers in epistemology. The paper gave rise to the notion of the "Gettier problem" for the otherwise appealing view that knowledge is justified true belief. A great deal of epistemology since can be understood as trying to figure out what to add to the "justified true belief" theory to get around Gettier-style objections.

Also interesting, it seems to have been basically the only thing he published, and he published it mainly because he had to publish something.

A more interesting zero-word article than the writer's block one is "On Nonrecoverable Deletion in Syntax" (1972) [0]

0: https://pbs.twimg.com/media/BmLQjz0CMAA-l28.jpg:large

A nice example of short papers that introduced very important concepts are C.A.R. Hoarse publications on the partition and quicksort algorithms in 1961: * https://dl.acm.org/citation.cfm?doid=366622.366644 * https://dl.acm.org/citation.cfm?id=366642

Though an interesting article, it should be noted that the article reproduces and makes available copyrighted - and expensive - materials. Case in point, a $40 PDF that was converted to a .png and displayed on the site:


To be fair, I should note that those publications published only the abstract and eliminated the actual research results.

(lrf v'z wbxvat)

Reminds me of the classic, "The Effects of Peanut Butter on the Rotation of the Earth" from the Annals of Improbable Research.


Another short, important paper:

E. W. Dijkstra, “Solution of a problem in concurrent programming control,” Commun. ACM, vol. 8, no. 9, p. 569, Sep. 1965.

Just one page, no references (understandable, since it is one of the very first papers on concurrent algorithms), proposes a mechanism for mutual exclusion between N processes.

Was the superluminal neutrino paper still relevant after they discovered it was an error?

Or it's something completely different?

Negative results are pretty useful. Eg the next time someone comes across a similar measurement, they can take that paper for some inspiration of a list of things to check.

Has anyone made a self referential paper on recursion? That would tickle me pink.

Is there any paper that can be both short, concise and deep?






Ducks swim.

E=mc^2 . Thank you and good night.

What’s purpose? Seriously?

"The unsuccessful self-treatment of a case of “writer's block”1" by Dennis Upper has a total of zero words.


Did you read the whole article? This example and a couple other "empty page" examples are given lower down in the list.

From the HN guidelines:

Please don't insinuate that someone hasn't read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that."

> Please don't insinuate

He or she didn't. He or she was quite direct. See https://en.oxforddictionaries.com/definition/insinuate.

If anything, it is you who are insinuating that the commenter did not read the HN guidelines. The guidelines misuse the word 'insinuate' to mean 'say or insinuate' or possibly 'accuse'.

You are correct, thank you for pointing it out, I'll try to be more conscious about it in the future :)

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact