5 ms·

I've done a lot of work on tools that interface with the academic publishing infrastructure. Elsevier not only make a huge profit margin, but they operate at a

by Blahah 10y ago

I've done a lot of work on tools that interface with the academic publishing infrastructure. Elsevier not only make a huge profit margin, but they operate at a staggering level of inefficiency. They are haemorrhaging money and producing precisely nothing of value, and still making profit for their investors hand over fist. Quite incredible.

chris_wot 10y ago
Can you give examples of the inefficiencies you discovered?
secfirstmd 10y ago
Hmmm...it seems like the space is ready to be disrupted. How might that look? I mean purely publishing isnt just ideal as it lacks the so-called quality control and prestige of established publications. What might be a middle ground for disrupting this crappy monopoly on the world's knowledge?
- zazen 10y ago
  Academics in some fields have been starting independent open-access journals, sometimes with the explicit goal of replacing Elsevier journals in particular. https://gowers.wordpress.com/2013/01/28/the-elsevier-boycott-one-year-on/ https://gowers.wordpress.com/2013/01/28/the-elsevier-boycott...
- Blahah 10y ago
  Personally, I think it has at least these important aspects: - built on a foundation of open source software - new ways of presenting and consuming should be encouraged and easily discoverable - completely outside the influence of the current publishing industry There's a community of us working on this stuff. We hang out on gitter (https://gitter.im/codeforscience/community https://gitter.im/codeforscience/community).
- VMG 10y ago
  As long as you have government power (public funding) in academia, you'll have rent-seeking.
- fossuser 10y ago
  Need to figure out how to align the incentives and fix the prestige problem of academia in order for this to work. Open Access exists, but you don't get the status you need from it (and weirdly it often costs a couple hundred dollars for the graduate student to publish their paper open access in addition to the Journal). It'd be nice if there was a way to make the community more like open source software where you get status from sharing and collaborating instead of from hoarding information in fear of being 'scooped'. As it stands the incentives are directly opposite collaborating and sharing information and really the best parts about science and research.
  - return0 10y ago
    Bingo. But unfortunately life scientists stubborny refuse to think like programmers.
- return0 10y ago
  You would need a shift away from publishing i think. The rationale behind publishing in long-form journal articles is not really valid in our century. There is an immense amount of noise in each paper, and findings are typically summarized in their 3-4 figures anyway. There are proposals for new approaches (e.g. http://www.silvalab.com/silvapapers/QAAlcinoSilva.pdf http://www.silvalab.com/silvapapers/QAAlcinoSilva.pdf). I believe we need a directory of scientific findings of sorts. It should be possible for each finding to cite other findings it relied on, thus creating a graph of influence/significance. The current situation with citations often involves friends citing friends , so i don't consider it reliable. It should also include open questions , thus giving research directions.
  - throwawaybookst 10y ago
    Out of curiosity, have you spent time in academia? > The rationale behind publishing in long-form journal articles is not really valid in our century. No, it's perfectly valid. Some ideas can't be expressed in short form. A 5-page long proof requires a 10-20 page article to properly motivate and provide context/explanation. An evolutionary or piece-by-piece doesn't make any sense; you'll end up with fragmented dead ends. Waste of everyone's time. > and findings are typically summarized in their 3-4 figures anyway I'm not really sure which fields you're referring to, but I assure you that you're over-generalizing. This isn't even remotely true even for very empirical sciences. > There are proposals for new approaches Usually these are appropriate only for a single subdomain or methodology and only provide one particular and extremely opinionated view on the results. Such as the proposal in the article you posted. The overhead isn't worth it and there's a huge risk the relevant field(s) move quickly enough that the presentation method becomes obsolete before it becomes useful. > It should be possible for each finding to cite other findings it relied on ...I don't know what to say to this... > thus creating a graph of influence/significance Yeah, we have this. It's called bibliometrics. UNIVERSALLY HATED by anyone who's not a bean counter. Good people ignore them. Bad people optimize for the metrics and it becomes a stupid game that has nothing to do with the thing you're trying to measure. > The current situation with citations often involves friends citing friends , so i don't consider it reliable This is just one reason among many that bibliometrics (a.k.a. any necessarily poor attempt at "a graph of influence/significance") are a poor mechanism by which to judge science... > It should also include open questions , thus giving research directions. In many fields literally every paper includes this. In most it's rather obvious to anyone who comprehends the paper what the next steps are.

by Blahah 10y ago

chris_wot 10y ago

Can you give examples of the inefficiencies you discovered?

secfirstmd 10y ago

Hmmm...it seems like the space is ready to be disrupted. How might that look? I mean purely publishing isnt just ideal as it lacks the so-called quality control and prestige of established publications. What might be a middle ground for disrupting this crappy monopoly on the world's knowledge?

zazen 10y ago
Academics in some fields have been starting independent open-access journals, sometimes with the explicit goal of replacing Elsevier journals in particular. https://gowers.wordpress.com/2013/01/28/the-elsevier-boycott-one-year-on/ https://gowers.wordpress.com/2013/01/28/the-elsevier-boycott...
Blahah 10y ago
Personally, I think it has at least these important aspects: - built on a foundation of open source software - new ways of presenting and consuming should be encouraged and easily discoverable - completely outside the influence of the current publishing industry There's a community of us working on this stuff. We hang out on gitter (https://gitter.im/codeforscience/community https://gitter.im/codeforscience/community).
VMG 10y ago
As long as you have government power (public funding) in academia, you'll have rent-seeking.
fossuser 10y ago
Need to figure out how to align the incentives and fix the prestige problem of academia in order for this to work. Open Access exists, but you don't get the status you need from it (and weirdly it often costs a couple hundred dollars for the graduate student to publish their paper open access in addition to the Journal). It'd be nice if there was a way to make the community more like open source software where you get status from sharing and collaborating instead of from hoarding information in fear of being 'scooped'. As it stands the incentives are directly opposite collaborating and sharing information and really the best parts about science and research.
- return0 10y ago
  Bingo. But unfortunately life scientists stubborny refuse to think like programmers.
return0 10y ago
You would need a shift away from publishing i think. The rationale behind publishing in long-form journal articles is not really valid in our century. There is an immense amount of noise in each paper, and findings are typically summarized in their 3-4 figures anyway. There are proposals for new approaches (e.g. http://www.silvalab.com/silvapapers/QAAlcinoSilva.pdf http://www.silvalab.com/silvapapers/QAAlcinoSilva.pdf). I believe we need a directory of scientific findings of sorts. It should be possible for each finding to cite other findings it relied on, thus creating a graph of influence/significance. The current situation with citations often involves friends citing friends , so i don't consider it reliable. It should also include open questions , thus giving research directions.
- throwawaybookst 10y ago
  Out of curiosity, have you spent time in academia? > The rationale behind publishing in long-form journal articles is not really valid in our century. No, it's perfectly valid. Some ideas can't be expressed in short form. A 5-page long proof requires a 10-20 page article to properly motivate and provide context/explanation. An evolutionary or piece-by-piece doesn't make any sense; you'll end up with fragmented dead ends. Waste of everyone's time. > and findings are typically summarized in their 3-4 figures anyway I'm not really sure which fields you're referring to, but I assure you that you're over-generalizing. This isn't even remotely true even for very empirical sciences. > There are proposals for new approaches Usually these are appropriate only for a single subdomain or methodology and only provide one particular and extremely opinionated view on the results. Such as the proposal in the article you posted. The overhead isn't worth it and there's a huge risk the relevant field(s) move quickly enough that the presentation method becomes obsolete before it becomes useful. > It should be possible for each finding to cite other findings it relied on ...I don't know what to say to this... > thus creating a graph of influence/significance Yeah, we have this. It's called bibliometrics. UNIVERSALLY HATED by anyone who's not a bean counter. Good people ignore them. Bad people optimize for the metrics and it becomes a stupid game that has nothing to do with the thing you're trying to measure. > The current situation with citations often involves friends citing friends , so i don't consider it reliable This is just one reason among many that bibliometrics (a.k.a. any necessarily poor attempt at "a graph of influence/significance") are a poor mechanism by which to judge science... > It should also include open questions , thus giving research directions. In many fields literally every paper includes this. In most it's rather obvious to anyone who comprehends the paper what the next steps are.