Slashdot Mirror
Tipping Point For Open Access CS Research?
First time accepted submitter trombonehero writes "Prominent Computer Science researchers from Google, Microsoft and UC Berkeley are starting to sign the 'Research Without Walls' pledge, promising to never be involved in peer review for a venue that does not make publications available to the public for free. Others have made similar pledges in isolation; could this be the start of something big?"
Hard to believe it can ever really happen. But whatever, I know I am a cynic.
"Ok, let people learn and use our patented ideas so we can sue them later."
This is all well and good, but considering how much money is getting thrown around for research and publication and patents, I wouldn't be shocked to see some organizations that require a fee for publication access go straight to universities and the like with a "Hey, your professors need to stop going to these completely open venues, or we're going to stop publishing every other professor you have!"
Sadly, I don't think academia is pure enough to make this work.
Between the 1930s to 1980s, most of the important computer science research was done. Since the 1990s, however, much of it has been utter rubbish. This goes far beyond the general "95% of everything is shit". There is basically no good research happening today, anywhere.
Hell, just look at any list of recently published papers. Many of them will just be buzzword-laden crapfests. They throw in as many buzzwords as possible, likely to try to get funding from industry vendors.
Of the remaining papers, those dealing with programming languages will merely be the researcher or researchers rediscovering some fact or property that we already observed back in LISP in the 1960s.
Papers dealing with algorithms are often merely minor tweaks to existing algorithms these days. There hasn't been anything truly groundbreaking in this field for decades.
Those dealing with networking are often proposing some new protocol to do something we've been able to do for ages. Worse, they're often merely tunneling one protocol through another and calling it a breakthrough. These days it's usually "some-existing-protocol over HTTP".
Then there are those bullshit research papers that propose some useless "model" of some over-simplified system. These are perhaps the worst of all. Hell, a bunch of interns in industry can often come up with more useful and realistic models and simulations during their summer placements.
I'm very thankful that I was able to be a researcher in the field when there were real discoveries getting made, and at least had the opportunity to retire before things got too shitty.
The free market would correctly price and maximize the speed at which research is done. It has been proven by history and is under attack by the modern socialist hippie culture. The one thing we could do to help is to reduce the amount of patent trolling.
[C]ould this be the start of something big?
Call me when you get medical researchers to sign up for something like this. CS is a small backwater that the general public (and other fields, frankly) will not notice.
It's a good thing, but not necessarily earth-shattering. It would be nice to see articles out from behind the IEEE and ACM paywalls, though.
That is all.
CS people are really good at posting all their work on their personal pages, offering preprints and reprints for free (heck, CS people even post their research statements and tenure documents, their family photos and their book reviews!). So making this official would be a nice statement, but not a significant shift in practice...
As of this week, ACM authors can now post copies on their personal/institution website, through a somewhat convoluted process. I think ACM already let you do this (I've never had any cease and desists in any case), but this is now integrated into their digital library and contributes towards your download count. A small step, but a good step.
This is big. There are a lot of parasitic journals today, that is, journals that take work the public paid for, and lock it behind paywalls. Parasitic journals typically use big-name free labor from to do the peer reviews. If the world removes from the parasites many good peer reviewers, as well as many good papers (through policies like the NIH Public Access policy), then they will have to change or fold. I don't have a problem with organizations paying for work to be done, and then charging for use of the result. For example, most fiction authors get at least some money for their labor (not a lot, but at least some). In contrast, parasitic journals typically take publicly-funded stuff away from the public; time to change. By the way, there are a number of journals and publications that have always done this, like ACSAC. If authors would simply ONLY submit their works to open access journals and publications, the parasites would disappear.
- David A. Wheeler (see my Secure Programming HOWTO)
And the one that matter for me (a researcher) is how I get funded. Basically I get funded when I can convince other people of how good I am. To estimate that, they look WHERE I am publishing my research; and most likely, they do not look at WHAT I am saying. The name of the conference or the journal is what matters most. What you are actually doing is not so important.
I know that suck. It makes me cry at night. But that is what it is. If I came not to publish in journal with no public open access, I won't be able to publish in journals that matter in my field. So I won't get funded.
I totally agree the public should be able to read what ever we write. But I can not give up my funding. (For the record: no funding, no food on my table.)
Moreover, that's basically a false issue. All journals and conference allow you to publish pre-print on your website. All my papers are on my website or in arxiv. So I am not even sure it matters so much.
Of course, complete open access for everybody would be better.
My biggest problem is wading through the crap publications to find the good ones. 20 years ago we called publishing incremental results "salami" science. Oh goody your arbitrary change to an algorithm improved it on some meager special case test set, and your publication is so short in the intro and discussion that you don't link it to the extisting stream of knowledge or compare it broadly. Your publication is unlikely to have any value except to adding length to your resume. Otherwise is has almost negative value since it's indistinguishable from all the other ones like it that are actually wrong.
Publishers at as a bit of a gate keeper in this tragedy of the commons. Publish less not more.
Some drink at the fountain of knowledge. Others just gargle.
what the fuck are you talking about. "small backwater"? eat shit.
Easy tiger. If you cherish an academic discipline so much, one would imagine that you could bring your point across in a more articulated manner. I hardly believe an academic discipline needs (or appreciate) 3-grade retorts, specially if it is an academic discipline that started as a branch of Mathematics, and in which Mathematical Logic plays an important role.
He does have a point in that CS is a very small discipline in terms of its body of knowledge (in relation to other STEM fields). CS by itself is just short of 4 decades old, and the study of computability barely a century old. This opposite with the other engineering and science fields that have hundreds of years, on top of millenia of study and practice.
Don't confuse the pervasiveness of computing in modern life with the pervasiveness of Computer Science proper among the other STEM fields.
I do disagree with him, however, on his assessment that this is or might not be earth-shattering. For one, the fact that is occurring at all is earth shattering. And secondly, most earth-shattering events in science and academia do not start with a bang out of nothing (instead, they start from seed events and findings that gain momentum over time.)
Furthermore, even if CS is small fish when it comes to academic research, the institutions (both academic and private) that are pushing for this aren't small fish themselves. There is enough muscle there to bend arms.
I totally agree and you can see that many people have the same problem. For example, there's things like "Faculty of 1000" which are sort of a cross between peer review and social networking. Basically, important people tell you what articles they read and liked. It's not open to everyone's vote. This helps focus this sprawling mess.
I only read the very highest impact factor journals now. I only read other journals when they are linked from an article in a high impact journal.
One needs more filtering not less. If publishing is free, it leads to overgrazing. Entry fees help everyone including the person publishing. Moreover the fact that they might be prohibitive to an individual but not to an institution is a good thing. It helps science advancements spread when people work in teams not as lone wolfs, no matter how brilliant.
Open publishing may sound good but it's not something I will be reading.
It's a win-win situation. Guilds (which is exactly what many pay-for-access journals function as) eventually starve or die for a reason: given enough time they're eventually self-defeating, just like any other kind of hoarding/monopolization taken to extremes.
NO
This is what politics looks like now
I don't know why the second link is really upset about the preprint policy. In fact, not long ago being allowed to put preprints on a public server was considered a victory.
For those who don't know, a preprint is not the version you submitted to the journal, but the version just prior to publication: After the peer review, after the formatting, and after all corrections. It's the one they send to the author saying, "This is how your paper will look - do a a quick glance to see if you find any errors." It's almost always identical to the final paper that is published.
So if they allow preprints, then it's as good as the final paper. To complain that they don't allow the final paper is just whining.
Beetle B.
It's a start.
You wouldn't just allow mass audiences to review, you'd Meta-Rank the Reviewers. So if Dr. ______ reviewed something well, and most of his reviews are insightful, you just "follow him" or something. That's what the Peer Review is like in the normal course. "Is this article sane? I am staking a small portion of my rep giving my Yay or Nay."
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
So you'd advocate non-anonymous referees?
Beetle B.
It's likely that many people here are sick of all the expensive academic periodicals that often contain interesting articles, but that almost no one without access to a university library can read... except perhaps for a summary. One of my interests is herpetology, which is pretty obscure, but nevertheless, there seem to be hundreds of periodicals published on this one very narrow subject alone. I've also heard stories about researchers who were upset to find out that their own papers, once published and on which they worked so hard, turned out not to be publicly available anyway. At that point the publishing company owns the copyright on your work, so AFAIK you can't even publish it on your own website.
Once upon a time, I suppose getting your paper published in one of these obscure and specialized academic periodicals was one of the only options available. Every few months a limited number would be printed and mailed out to university libraries around the world, but only to the ones with an interest in the subject (and willing to pay the subscription fee). Even today, however, the number of academic periodicals out there seems way out of proportion, so how come?
Well, it turns out that everybody would like to publish in the better known periodicals, but that most papers are rejected for various reasons. Still, as an academic you have to keep publishing (publish or perish!), so then you try to get your work published in one of the lesser known ones, perhaps with less peer review. This is one explanation. Another is that corporations often like sponsoring periodicals; sometimes just a little, and sometimes so much that they launch their own as a vehicle with which to advertise their products. However, peer review usually isn't worth all that much in cases of the latter.
The times, however, have now changed things rather dramatically. With the Internet, in my mind there is little or no need for these wretchedly expensive periodicals any more. Are they really good for science anymore, or are they only good for generating revenue for the publishing companies? Actually, it may be that things will change regardless. Of the periodicals that depend on sponsorship to some degree, many of their sponsors are now losing interest, probably because they think their activities on the Internet are more important. It's true that papers still needs to get peer-reviewed, but if it turns out that most researchers are now willing to do that even when the publications will be made available to the public for free, then everybody wins, right?
Perhaps I should add a note regarding medical periodicals. From what I understand, whether these are associated with expensive subscription fees or not, access to them is often restricted. If so, this is because it is felt that if the general public (e.g. journalists) had access to them, that the articles would more likely be misinterpreted. However, since this is bound to happen anyway, and it's worse if most people only have access to a misinterpretation of the article, I see no excuse for this practice either.
And universities having to pay extortion money ($5,000 per year in some cases for a single journal) is also a non-issue?
You are simply describing as inevitable the mechanism by which parasites exploit your work, so you are just looking for the quick way out. Shortsighted.
Not a false issue.
It has become increasingly apparent that the value of education received in upper crust undergraduate programs or post graduate degrees rarely equates to the dollars spent. What people are really paying for is a premium rung on the social graph.
It's rarely clear how one recovers the investment, unless you're one of the bright lights that launch onto a lucrative career track. How many humanities graduates ever see tenure? I saw a study the other day which determined that out of an especially strong draft year in Ontario (long ago), of the 30,000 kids enrolled in hockey at the lowest levels, two players became well-known NHL regulars, and maybe twenty got a cup of coffee.
But people continue to sign up for big debt and small hope, more out of fear of losing your rung than a clear idea of how the loan is eventually retired.
I'm wondering if soon every good job out there is obtained after ten gruelling episodes on The Apprentice.
Perhaps we can train Watson to assess a little more importance to WHAT and a little less importance to WHERE, so that the parallel aristocracy of competence doesn't die out completely.
It would be funny, you know, if the computers are the first intelligent life form to curl their lip at the social graph.
I would be much more impressed with someone making a pledge to never peer review any article which doesn't include (at least in the appendix) the experimental data on which the article is based. Most current publications do not have the requirement to publish the data. Which makes all publications nothing but inherently unreliable claims.
Any guest worker system is indistinguishable from indentured servitude.
Call him Dr. Snickerdoodle if you want. It doesn't matter what his name is if he's a good reviewer.
404: sig not found.
This is a great start, but doesn't go far enough.
>promising to never be involved in peer review for a venue that does not make publications available to the public for free
If you want to publish in leading Bioinformatic journals, you must make your code freely available.
I don't think you necessarily have to make the source code public -- but at least the executables.
Because, um, "science" implies your results should be reproducible.
It would be an interesting project (if someone were interested in funding it) to compare the
requirements and response times for the many (many!) computer science journals.
Strawman questions:
1. what percent make their articles freely available?
2. what percent require code or executables to be publically available as a condition of publication?
(this assumes the article is reporting results derived from working code, as opposed to
theoretical results)
3. what is the average turn-around time from submission to getting feedback
from peer reviewers, and (if accepted) until publication?
And the results could be broken down by sub-discipline: bioinformatics; IT; high performance
scientific computing; etc, etc.
since I'm posting anonymously, with some trepidation I'll post ;) email: dhysom2@gmail.com
my (secondary
Off topic, in my experience, the peer review results are back to you within six weeks or less.
Very different from other areas, such as High Performance Scientific Computing, where
you may not hear anything for a year or more.
The problem here is that someone needs to organize these things.
True, but not relevant. The necessary organization and infrastructure can be done quite cheaply, if organizations are willing to move from the past to the present.
Someone has to pay for the bandwidth, buy and run the servers, spend effort soliciting reviewers, run the reviewing software, respond to questions, request ISBNs, submit the work to indexing sites, etc etc etc..
I think you're grossly overestimating the costs if they switch to an exclusively digital realm, which is where they need to go. Bandwidth and running basic servers (which is all that's needed) are commodities. WebHostGiant will provide bandwidth and basic servers for $2.79/month with no bandwidth maximum. A serious journal will want more than that, yes, and there are other requirements too. For example, you'll need DNS registration (which companies like GoDaddy or 1and1 will do for cheap). But really, these are highly-competed commodities, so we're talking about petty cash budgets in most companies. You could probably set up the technical infrastructure for less than $100/year, plus 20 hours of volunteer time for the initial setup. Just get 5 people to contribute $20/year and you're done. The exact number isn't important; what's important is that it is REALLY REALLY small compared to paper-based publication.
Yes, you need some code to manage reviews, but not much. Existing programs can do the job; you really don't need more than a Wiki or mailing list. You can get some efficiencies using specialized programs, but their development could be amortized across lots of different journals (it's perfect for creating as open source software; there are probably several such programs already). It's clearly possible; Wikipedia's MediaWiki, for examples, wasn't developed by a for-profit organization, and it handles scales far beyond what journals require. You need an ISSN, not an ISBN, for a magazine, and that's a one-time expense.
The "soliciting reviewers" and so on can be done by volunteers, as is already often done.
The basic problem is that paper publishers haven't noticed that the digital economy is VASTLY cheaper (or they're scared BECAUSE they realize this).
Open access journals address this by having a publication fee, or by advertising, or by seeking volunteers and asking for charity. Paying for publication gets blurry with vanity press, and advertising ends up sucking up to the advertisers.
So stop pushing the paper. Push the bits, and let people print what they want. For academic works, what people (should) want is the information, not the pretty binding.
Volunteering and charity seem wonderful, but have to compete with lots of other worthy causes.
Absolutely true. But if it's easy, it's not too bad. Also, volunteering for journals is considered something valuable in many academic circles, so it's not just charity in many cases. The paywalled journals depend on volunteerism already, so it's not a change in that respect. And when people don't feel exploited, they're more likely to volunteer.
In CS, most of the major publishers allow you to post a copy of your paper on your personal website (as long as you also link to the official version). Finding papers outside of the paywalls is only difficult when the authors are in industry (but those aren't the publicly funded papers you're talking about anyway).
Fair enough. But typically CS researchers are writing papers because they want to get information out to the widest possible audience that would be interested. Paywalls just get in the way, for reasons that are now obsolete.
- David A. Wheeler (see my Secure Programming HOWTO)
by signing the Budapest Open Access Initiative, supported by Soros' Open Society Institute. You can read about it here http://www.soros.org/openaccess/read
Only a few researchers in each field need publish the findings of their field in order to quash the information hiding and distribution restricted activity which has created a "monopoly of knowledge". The players include the mega -corporate global feudal lords, the universities seeking to profit from government research findings, the insider publishers, the law makers, etc. This lock up the knowledge of science program behind closed high priced right person gated doors of copyright has transformed our supply/demand economy capitalism into supply side only feudalism. Instead of serendipity generating competition, our government, by its laws, has turnicated competition and create global mega sized corporate feudal estates and transformed those non monopoly holders into serfs if they work for the global mega sized corporate feudal estates or paupers if they don't. Remember monopoly is a product of the rule of law. Monopoly cannot exist without a government, commerce corrupted government officials and the use of the government's force to impose its anti-competitive laws against would be commercial competitors. Copyright is being used to terminate or severely limit the distribution of scientific knowledge unless the user has great big pockets and this condition plays a big part for the accumulation of inventories of tax-free monopolies enjoyed the few mega corporate feudal lords. The global mega sized corporate feudals use monopoly inventories, copyright, patent and license of parts of the public commons, to inhibit competition. Without competition there is no capitalism instead it is feudalism. The Internet would never have developed had it been subjected to the copyright and patent monopolies current law has imposed. Its impact is felt everywhere in the development of equipment, in the development of software, in gaining access to the public markets, in the learning environments, etc.. Everywhere monopoly exists it inhibits competition and without competition there is feudalism.
Two classes of reviewers.
This is a story about professional science, so already they only give the article to other professionals for the Peer review. Whether or not those Peers are revealed (intentionally - we're discussing process here) is a process choice with good and bad. In the first class at least a Senior Board needs to know that the Peers are qualified scientists. Quick example is the (GM) tag for chess grandmasters on ICC. Except for flukes, that tag doesn't get there by accident, but the real name is often private because they want to be left alone from being pestered.
But we need a class of "Intelligent Laypeople", who offer quality comments, even if they make mistakes. That's like a High-Karma user here. (Again, leave out the edge cases of "Karma Experts.)
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine