Let me give you an unpopular answer -

- outside top 500 scientists across all the fields combined no one produces quality papers.

- the manuscripts received by EICs are garbage. EICs themselves rarely touch anything, or edit anything.

- "peer review" process in all non-the-most-upper-tier publication is a joke.

- copy-editing and proofreading is non-existent.

About 15 years ago a lot of journals that are now owned by the major publishers were society publications. This means that societies owned the title, owned the copyright and owned the process. Societies could not find people capable and willing to run a journal for $5k or so a paper including distribution and printing. So the societies went to the likes of Springer, Informa, T&F, Elsevier etc. Those said "Ok, sure. $2k per paper and we produce and distribute it or $5k per paper and we do everything". That was too much for societies. Instead the societies said said "Hey, what if we sold you the rights to the journal? Would you in that case do all of it for free, and let our EIC suck his or her thumb while still being the man/woman on the title page?" "Of course" said the publishers.

And so now we are here. Production costs did not go down. Distribution costs did not go down. In fact, distribution costs increased because before that the journals were printed by a little printers in the US as the volume of most of journals is microscopic ( later printing went to China in the same microscopic quantities ). Now, however, journals also need to be distributed in all kinds of wacky XML formats on all stages of production, which needs to be coded, in most cases by hand.

So unless every Joe, Jack and Jill, the scientist, wants to go back into the publishing world nothing is going to change.

[Source: pillow talk]

> outside top 500 scientists across all the fields combined no one produces quality papers

Just to make sure I read this correctly: Are you saying that, when judging by paper quality, there are only 500 good scientists on Earth?

Only about 500 produce readable manuscripts to start. Publishers have their names because it changes schedules that much.

That sounds absurd. What are your sources?

Pillow talk with someone who ran production for hundreds of STEM journals for two of the publishers

A root cause there would be the opportunity cost the scientist/society calculates to publish themselves vs have a publisher do it.

The equation today isn't the same - with open source / community effort, I suspect the cost/effort of publication today doesn't even come close.

I was just randomly looking at what it would cost (rough estimate) for 25TB of S3 storage - 2,000,000 put requests a month, 2,000,000 get requests, 25TB bandwidth OUT and 25TB bandwidth IN - and it's less than $3,500 per month. So the cost of hosting the published item seems actually quite reasonable... reducing the cost of the rest of the "publication" plumbing are problems that technology also helps address.

That's not the cost. The cost is to take a garbage paper with broken figures, broken text, in a barely parse-able English, written by someone who does not believe in citations or attributions, does not follow Chicago and produce a paper that is being distributed.

Late addition: The cost of distribution of a scientific journal in electronic format is minuscule - put it on a blog. Medium.com is free.

Another addition: I have seen a few of the author contacts. All of the ones that I saw said the author could publish his original manuscript on the web to free. The author could just not publish the result of the publisher's processing of the manuscript. Someone should ask the authors why aren't they publishing their original submissions.

> Late addition: The cost of distribution of a scientific journal in electronic format is minuscule

"The annual budget for arXiv is approximately $826,000 for 2013 to 2017, funded jointly by Cornell University Library, the Simons Foundation (in both gift and challenge grant forms) and annual fee income from member institutions."

https://en.wikipedia.org/wiki/ArXiv

> Medium.com is free.

Until its not, or until, as a commercial enterprise, Medium decides to completely pivot, or decides a controversial paper is too much bad publicity. If I publish to Medium, will future researchers be able to find my article in 10 years? In 50? Also I don't see how you've taken peer review into account here.

Agreed that the best publishers act as both a filter and improve the production quality of the papers, and that this is a true cost. I'd hope that they could continue to do what they do and also get paid by their subscribers. Ideally, they could do that AND the works could be published for free, and those who value their efforts would in turn reward the publishers for their efforts as well by subscribing to the paid version of their publications.

That said, Wikipedia mostly works. Between professional editors and the community, decent knowledge is categorized, edited/peer-reviewed and interlinked. Why the same cannot be done with academic papers is beyond me, other than to assume people simply don't want to change the status quo.

I also agree with your last point - so long as authors can contractually publish their own work, they should - and the Internet should make it as easy as humanely possible. I've published a paper before to an academic journal and had that clause, and then put it online on my own site as well (until I took my site down).

Random thing - I do think the standards of what constitutes knowledge matter, but boy do we hold them so high. I think there are many forms knowledge comes in, and expecting them to all be polished perfection is perhaps too lofty a goal. Knowledge is always a work in progress. Discerning between fact and fiction is the same whether you have a polished paper or a garbage one, it's just one is a lot easier to consume and you have more trust inherent by the nature of the reputation associated with the publisher, author, citations, etc.

>outside top 500 scientists across all the fields combined no one produces quality papers.

This is an exceptionally incorrect statement to make. The other points are fine, which only makes it more baffling you would make this point.

I'd rather see that money put into actual research rather than paying for access to free things. And approximately none of those funds go to pay for research.

Agreed.. is that like saying that researchers should do a "pay what you want for my paper" online and self publish/host the knowledge? Maybe somebody just needs to provide a standard "pay what you want" feature that academics can embed easily into their own site, and then self-publish / host.

Do you think pay to cite would be a good model?

No, not at all. Nor would models like "pay into a pool and get paid based on citations", either. That'd create perverse incentives, not least of which more incentive to push for citations to your own works when doing review.

Sure, but if you're paid based on the number of people citing your work doesn't that create the right incentives? People making more money are precisely those providing valuable information/methods to other researchers.

Sure, there's an ethics problem during reviews to push additional citations for people in your university or something but that's not too different from the current situation with grant applications, just more direct.

Hmm, that's interesting - I think maybe that's just like licensing, right? Except in this case, the licensor is the researcher/author and the licensee is the author citing the source?

Combining the two thoughts, "suggested payment to cite" with a "pay what you want" model could interesting too.

I suppose this would be a direct transfer of value from those who want to build knowledge on other people's work, to those who did the work. As long as it was not unduly expensive to do so, maybe - after all, the saying goes "paying my dues". I would think you wouldn't want to inhibit or slow down people from building on knowledge either (which is what things like Sci-Hub seem to be trying to avoid).

I wrote a few published papers some time ago, and I was told that without citations, a paper would be rejected as academic research should always be based on enhancements or references to prior work. So it seems kind of built in to the mechanism that citations are the requirement to build knowledge on top of when creating such works.

It would be interesting to see how a model like this could work. Since the value created by the researcher is essentially part of a web of value, it might be hard to determine who should get credit.

For example, we have 3 existing papers, A, B and C, and one new paper, D. Paper A cites paper B, which cites paper C - and all of them are additive to our knowledge base. Therefore, as the author of paper D, if we were to pay to cite paper A, should that also include some payment to the authors of paper B and C too?

I suppose one addition you could add to this - payment to only LIVING researcher(s)/author(s). If the researcher/author is deceased, then the payment stops. In this regard, it sounds like royalty/licensing fees.

in a way researchers paying for reading research is an absurd inflation.

Therefore, should the results of all research be made free and public? I was hoping we could allow people to pay if they wanted but also get it for free, and have both be legitimate.

Big and famous universities have deals, but not the rest, which is the reason why scihub exists in the first place. Maybe publishers should up their game and provide near free access to any PhD student or higher and a better uptime (people laughed at the fact that scihub+libgen hack was more reliable than springerlink).

Students would pay a little fee, if that doesn't impede their research too much.

Also I was thinking of a time based value. Papers / Journals older than X years, or that have already received a fair amount of fees could fall into free access. That could make students looking for alternative or forgotten ideas more motivated.