Hacker News Reader

Comment by ksk | Hacker News Reader

>There are better ways to do science in a world with computers and the internet.

Well, propose an alternate system that stands a chance of being adopted then.

AI research is experimenting organically with ways to identify important papers (everything gets published on arxiv.org, but that's far from a replacement). A new system is more likely to emerge from all this experimentation than to be invented in whole.

Conferences play a large role, and automated tools like arxiv-sanity.com and Google Scholar (for mapping citations) help. Even Twitter is important, if a major figure tweets about a paper, many people will read it. Conferences are experimenting with models to reduce the lag time from submissions and trying different systems for selecting papers.

We'll see where it goes.

Well, all of those things you mentioned are voluntary, optional and not guaranteed. I think having people whose job it is to peer-review, be selective, be vigilant of bad-science, etc is a good thing. I'm completely for open-access, but I'm against having zero selection criteria. It puts a drag of everyone if they have to wade through thousands of junk papers while doing research.

If hacker news did not have a voting system/flagging system/banning system it would just be a giant cesspool of links and comments submitted by bots.

No one's job is to peer review for established journals. Peer review is unpaid.

To your point, yes, there are plenty of journals that don't pay (but pay for other activities). Perhaps I should have just called it a fee instead of a job. Here are some that do pay - http://academia.stackexchange.com/questions/20930/

Interesting link! I wasn't aware of anyone paying for peer review when I was in academia. It appears from the linked examples that it's still rare and relatively small dollar amounts, so indeed not a job.

I love the example of the New England Journal of Medicine finding that paying $5 was worse than not paying, as potential reviewers found it insulting.

As to other activities, yes, I think companies behind journals typically pay people to do administrative functions for the journal, and the editor may or may not be paid (for Nature, I believe the editor is paid). However, it is possible to replicate the traditional review structure in a completely free OA journal with volunteered resources; in computer graphics, I think JCGT does this: http://jcgt.org/about.html

>"It puts a drag of everyone if they have to wade through thousands of junk papers while doing research."

If something that sounds interesting was published in Nature/Science around 2000, I will usually just skip it. I already know the methods section will be crap and I will be unable to figure out what process generated the data (at least in the case of biomed). When I needed to dig through that stuff and try to make sense of what was going on in detail, that was an awful nightmare. And that was an editorial policy... now they at least let authors put the required info into supplements.

I think that was the era of "peak peer review", the idea has been losing clout ever since (mostly due to the internet).

If it has to be adopted, then it has to be based of that all science is supported by tax grants. It simply has to be accepted that all science that receives any government money has to be complete open and that using any portion of the grant to pay for this should be illegal.

Now that nobody (or close to nobody) can publish in Nature some other open access journal will be the number one.

Without the need for printing papers the amount of people actually needed to run a journal drops to a few admin people (research and peer review is still paid for by others) and somebody to keep the website running. This can easily be paid for by a few universities (Harward is sitting on billions and they have an interest in solving this problem).

Reply on news.ycombinator.com