4 ms·

If the number of citations is a measurement for achievement then perhaps we should put more value to the Like buttons we so frequently use. But, I'm sure he's g

If the number of citations is a measurement for achievement then perhaps we should put more value to the Like buttons we so frequently use. But, I'm sure he's good in what he does since so many people agree on his thoughts.

annadane 8y ago
...stupid.
diegoperini 8y ago
I understand the criticism but do you really think a like is the same thing as a citation? A like requires a single mouse click. A citation requires a published article.
- dschuetz 8y ago
  No, of course those are completely different things. The number still doesn't say whether his work is right or wrong. You can cite the Bible up and down and it still doesn't make it anywhere near plausible. As far as I know, Juan Maldacena is a supporter of String Theory, and despite of his minor achievements in science might being correct in context of his field, 17.000+ citations doesn't say anything about the quality of his original work on String Theory. So why even mention it? Just so to say that he's an authority in particle physics? If I had the same scientific authority as he does for a completely bonkers theory, I'd instantly say that the guy is nuts.
  - sgt101 8y ago
    I disagree - citations count is not a linear measure of quality of work, or importance, and there are significant anomalies and unfairness's in the way things get cited. For example, write an early text book and you may get a tranche of citations without doing a significant amount of important original research. Grad students who co-author with famous authors also get a significant boost due to the reputation and connectivity of their mentor. But, I think that most senior academics will have less than 15k citations in total. 17k citations for a single paper is remarkable; it's an indication of work that was important, original and definitive. I only referee in compsci, but in compsci citing papers that are nothing to do with the work in hand, and citing them to generate authority for your arguments does not wash and leads to rejections. Citations point to the basis of your work; establishing that you are building on firm foundations and differentiating your contribution from previous contributions to enable the reader to learn. Being cited in this way is being a teacher to the community.
    - dschuetz 8y ago
      You actually support my argument in the way that he's established himself as the teaching authority to the community. But imagine him being wrong and all of the people's work who cited his is wrong. All that knowledge and insights - wrong. Isn't it also an ideal of a scientist to remain critical of everything and just make assumptions that "if that is true then..." or "if this is wrong, then..."? So many people based their work on his ONE paper. Jeez, imagine how much it would cost to find out that this paper is worth nothing.
      - dschuetz 8y ago
        Here's a follow-up on why Mandacena could be wrong, and his work is not the only one: "The problem here is that of what is an “interesting region of theory space”. At this point the failures of string theory unification strongly indicate that it’s not such an interesting region. It seems likely that we’d be better off if most theorists focusing on phenomenology of this failed program were to pick something else to work on." And as far as I could follow his work and attempts to explain symmetry with economics analogies, I think my concerns are valid. https://www.math.columbia.edu/~woit/wordpress/?p=9375 https://www.math.columbia.edu/~woit/wordpress/?p=9375
        raattgift 8y ago
        Citations are not a "like"-like function; they are an admission by an author of a work that another work shaped its development. One must also cite a work if one is writing a paper which disagrees with that work's conclusions, criticizes its methods, or points out a fault in its theoretical content. This has a particular benefit for readers. Woit himself carefully cites works he strongly objects to in his book[1] that shares its name with the blog you linked to. - -- [1] Woit P. Not even wrong: The failure of string theory and the search for unity in physical law. Basic Books (AZ); 2006 Sep 4., pp 268-274. (cf. http://www.math.columbia.edu/~woit/NEWerrata.html http://www.math.columbia.edu/~woit/NEWerrata.html )
      - sgt101 8y ago
        Every scientist I know bursts with enthusiasm and excitement when there is even a hint that a well known and important result is wrong. It would literally be a bonanza for them, 99.9% of the time the excitement is misplaced, but scientists live for the 0.1% I would say that finding out that this paper is worth nothing becomes more expensive and more valuable with every citation - every time people cite work they are looking at it.
_Microft 8y ago
You better view papers as tools, facilitating further research and a citation as use of a tool. So a lot of citations indicate that a paper is a pretty useful tool for other people's work.
- dschuetz 8y ago
  Yes, I understand that. But if a fool buys a broken tool and builds his home with it, how long do you think this fool's home will last?
  - ithkuil 8y ago
    Indeed the number of deep/transitive citations would perhaps be an even more interesting measure
    - dschuetz 8y ago
      Is there any instance in science that keeps track of citation lineage of a theory or experimental results? It is an interesting question.
    - gus_massa 8y ago
      Usually with so many citations doesn't mean that there are 1700 independent citations in papers, only 1700 different papers. His paper is like a root of the deep citation tree. Many people cite their last papers, some recent papers in the area and the foundational papers in the area. So once you get a big foundational paper, you pile up citations for years. It would be a nice project to get the ADG of citations with this paper as root and make a graph of it. My guess it that it will be something that looks similar to tree (each paper in a branch would cite a few recent papers in the same branch), but most of the papers in the tree would cite also the root paper too.
globuous 8y ago
Actually, that's how comments and links are ranked on this very website. Moreover, still on this very website, not everyone can downvote, and the threshold is a quantitative measure of likes. So literally, you comment's font color is due to what you sarcastically suggest we do. Finally, i'm guilty of occasionally looking at some HN user profile and just to see how many upvotes it has. If the user has 15k+, I always go "wooooow" ^^ So yeah. But then again, as a fellow commenter has already pointed out, a like requires a click and a citation a publication.
- mercer 8y ago
  Correct me if I'm wrong, but assuming that someone's karma is comment karma + post karma, the number is pretty much meaningless. (and if that's the case, is there any way to separate the two?)
joppy 8y ago
The number of citations is not a measure of quality (does anyone really think so?). Having 17,000 citations means that in 17,000 papers, people found some reason to cite that work, meaning that it had a massive overall impact. Terrible papers can have an impact too, but anything that has that much of an impact is worth reading.
mayankkaizen 8y ago
There has to be some difference between liking some random Taylor Swift video on Facebook and citing a high energy physics research paper.
throwawaymath 8y ago
Just to be clear here, you're drawing a comparison between citations in peer reviewed papers - written and refereed by experts in a field - and Like buttons clicked by non-experts on social media? I feel compelled to ask if you're arguing in good faith, because if you have a valid point I can't immediately think of a worse comparison to use. Yes, obviously citations are a heuristic for achievement and Likes are not. Citations signal that one piece of research has had impact on another. It doesn't immediately declare the research is true and valid, but it does signify the research has held up to the scrutiny of n other researchers who felt it worthwhile enough to include in their own work. Likes do not involve any of that reputation staking mechanism, nor do they involve experts in a review process that can reject papers by design. It would be incoherent to claim 17,000 upvotes on reddit give a paper anything resembling the weight of 17,000 citations.
- wallace_f 8y ago
  I wonder if there is a good way to study biases in academia, which is maybe what the poster was getting at? Feynman criticized aspects of academia as being "rotten" for the record...
dang 8y ago
We detached this subthread from https://news.ycombinator.com/item?id=18520989 https://news.ycombinator.com/item?id=18520989 and marked it off-topic.

annadane 8y ago

...stupid.

diegoperini 8y ago

I understand the criticism but do you really think a like is the same thing as a citation? A like requires a single mouse click. A citation requires a published article.

dschuetz 8y ago
No, of course those are completely different things. The number still doesn't say whether his work is right or wrong. You can cite the Bible up and down and it still doesn't make it anywhere near plausible. As far as I know, Juan Maldacena is a supporter of String Theory, and despite of his minor achievements in science might being correct in context of his field, 17.000+ citations doesn't say anything about the quality of his original work on String Theory. So why even mention it? Just so to say that he's an authority in particle physics? If I had the same scientific authority as he does for a completely bonkers theory, I'd instantly say that the guy is nuts.
- sgt101 8y ago
  I disagree - citations count is not a linear measure of quality of work, or importance, and there are significant anomalies and unfairness's in the way things get cited. For example, write an early text book and you may get a tranche of citations without doing a significant amount of important original research. Grad students who co-author with famous authors also get a significant boost due to the reputation and connectivity of their mentor. But, I think that most senior academics will have less than 15k citations in total. 17k citations for a single paper is remarkable; it's an indication of work that was important, original and definitive. I only referee in compsci, but in compsci citing papers that are nothing to do with the work in hand, and citing them to generate authority for your arguments does not wash and leads to rejections. Citations point to the basis of your work; establishing that you are building on firm foundations and differentiating your contribution from previous contributions to enable the reader to learn. Being cited in this way is being a teacher to the community.
  - dschuetz 8y ago
    You actually support my argument in the way that he's established himself as the teaching authority to the community. But imagine him being wrong and all of the people's work who cited his is wrong. All that knowledge and insights - wrong. Isn't it also an ideal of a scientist to remain critical of everything and just make assumptions that "if that is true then..." or "if this is wrong, then..."? So many people based their work on his ONE paper. Jeez, imagine how much it would cost to find out that this paper is worth nothing.
    - dschuetz 8y ago
      Here's a follow-up on why Mandacena could be wrong, and his work is not the only one: "The problem here is that of what is an “interesting region of theory space”. At this point the failures of string theory unification strongly indicate that it’s not such an interesting region. It seems likely that we’d be better off if most theorists focusing on phenomenology of this failed program were to pick something else to work on." And as far as I could follow his work and attempts to explain symmetry with economics analogies, I think my concerns are valid. https://www.math.columbia.edu/~woit/wordpress/?p=9375 https://www.math.columbia.edu/~woit/wordpress/?p=9375
      - raattgift 8y ago
        Citations are not a "like"-like function; they are an admission by an author of a work that another work shaped its development. One must also cite a work if one is writing a paper which disagrees with that work's conclusions, criticizes its methods, or points out a fault in its theoretical content. This has a particular benefit for readers. Woit himself carefully cites works he strongly objects to in his book[1] that shares its name with the blog you linked to. - -- [1] Woit P. Not even wrong: The failure of string theory and the search for unity in physical law. Basic Books (AZ); 2006 Sep 4., pp 268-274. (cf. http://www.math.columbia.edu/~woit/NEWerrata.html http://www.math.columbia.edu/~woit/NEWerrata.html )
    - sgt101 8y ago
      Every scientist I know bursts with enthusiasm and excitement when there is even a hint that a well known and important result is wrong. It would literally be a bonanza for them, 99.9% of the time the excitement is misplaced, but scientists live for the 0.1% I would say that finding out that this paper is worth nothing becomes more expensive and more valuable with every citation - every time people cite work they are looking at it.

_Microft 8y ago

You better view papers as tools, facilitating further research and a citation as use of a tool. So a lot of citations indicate that a paper is a pretty useful tool for other people's work.

dschuetz 8y ago
Yes, I understand that. But if a fool buys a broken tool and builds his home with it, how long do you think this fool's home will last?
- ithkuil 8y ago
  Indeed the number of deep/transitive citations would perhaps be an even more interesting measure
  - dschuetz 8y ago
    Is there any instance in science that keeps track of citation lineage of a theory or experimental results? It is an interesting question.
  - gus_massa 8y ago
    Usually with so many citations doesn't mean that there are 1700 independent citations in papers, only 1700 different papers. His paper is like a root of the deep citation tree. Many people cite their last papers, some recent papers in the area and the foundational papers in the area. So once you get a big foundational paper, you pile up citations for years. It would be a nice project to get the ADG of citations with this paper as root and make a graph of it. My guess it that it will be something that looks similar to tree (each paper in a branch would cite a few recent papers in the same branch), but most of the papers in the tree would cite also the root paper too.

globuous 8y ago

Actually, that's how comments and links are ranked on this very website. Moreover, still on this very website, not everyone can downvote, and the threshold is a quantitative measure of likes. So literally, you comment's font color is due to what you sarcastically suggest we do. Finally, i'm guilty of occasionally looking at some HN user profile and just to see how many upvotes it has. If the user has 15k+, I always go "wooooow" ^^ So yeah. But then again, as a fellow commenter has already pointed out, a like requires a click and a citation a publication.

mercer 8y ago
Correct me if I'm wrong, but assuming that someone's karma is comment karma + post karma, the number is pretty much meaningless. (and if that's the case, is there any way to separate the two?)

joppy 8y ago

The number of citations is not a measure of quality (does anyone really think so?). Having 17,000 citations means that in 17,000 papers, people found some reason to cite that work, meaning that it had a massive overall impact. Terrible papers can have an impact too, but anything that has that much of an impact is worth reading.

mayankkaizen 8y ago

There has to be some difference between liking some random Taylor Swift video on Facebook and citing a high energy physics research paper.

throwawaymath 8y ago

Just to be clear here, you're drawing a comparison between citations in peer reviewed papers - written and refereed by experts in a field - and Like buttons clicked by non-experts on social media? I feel compelled to ask if you're arguing in good faith, because if you have a valid point I can't immediately think of a worse comparison to use. Yes, obviously citations are a heuristic for achievement and Likes are not. Citations signal that one piece of research has had impact on another. It doesn't immediately declare the research is true and valid, but it does signify the research has held up to the scrutiny of n other researchers who felt it worthwhile enough to include in their own work. Likes do not involve any of that reputation staking mechanism, nor do they involve experts in a review process that can reject papers by design. It would be incoherent to claim 17,000 upvotes on reddit give a paper anything resembling the weight of 17,000 citations.

wallace_f 8y ago
I wonder if there is a good way to study biases in academia, which is maybe what the poster was getting at? Feynman criticized aspects of academia as being "rotten" for the record...

dang 8y ago

We detached this subthread from https://news.ycombinator.com/item?id=18520989 https://news.ycombinator.com/item?id=18520989 and marked it off-topic.