because they were, in fact, computer generated gibberish.
Generate your own realistic looking computer science paper here, or, if you prefer pure math, you can generate a gibberish math paper here.
At least one of the math papers also seems to have been accepted by a journal (admittedly, not a reputable one).
Seems like this has proven the system works. A hundred and twenty papers were “written” and only one of them was accepted for publication - and that was only to a quasi-fake journal that charges people for publication. So it appears legitimate journals rejected all of these fake papers.
Indeed, what I get from this is the system works.
The articles were published in conference proceedings, which means they are supposed to have gone through some degree of peer review. I’m not sure where you guys are getting that they were withdrawn prior to publication.
I’ll admit I’m not familiar with scientific publication. But I assume there are various tiers involved. Some publication is presumably more “legitimate” than others. Other publication might be more akin to posting on a message board like this one. It’s publication in the legal sense and there are moderators who are reviewing the posts - but nobody’s claiming we’re published authors and the moderators are editing our work.
So where does conference proceedings fall on this spectrum? Is it the equivalent of a peer-reviewed journal or the equivalent of a wiki?
If I understand correctly, the papers were “peer-reviewed” and accepted for conferences, a form of publication. From personal experience I know the quality of research and review for some computer science conferences is dismal. (Anyway, by making multiple submissions from different institutions the authors may have been able to steer fellow conspirators to be the reviewers.) Are there are other disciplines with conferences of as dismal quality as computer science’s?
The paper mentioned, Rooter: A Methodology for the Typical Unification of Access Points and Redundancy, is worth a read by the board’s computer scientists. It will have you rolling on the floor in laughter. Even the bibliography is funny: “Hash tables considered harmful.”
If they’re conference proceedings, not real papers, it’s hardly cause for concern.
Another article cited in that biblio is “Moore’s Law considered harmful.”
Another: “Red-black trees no longer considered harmful.”
The Experimental Results section contains this piece:
I like how two of the authors on a paper about 802.11b are E. Schroedinger and E. Schroedinger.
No, over 120 of the computer science papers were published, although (AFAIK) only one of the math papers was. The published CS papers (well, those that they know about) have now been withdrawn from online databases, but they were out there and had supposedly passed peer review (though not for prestigious journals). Presumably there actually was no peer review, or only of the most perfunctory sort, but anyone finding them in the database would assume that they had been passed.
It all rather puts the Sokal affair into perspective, I think.
Isn’t it? I don’t know about computer science or math, but in the fields I do know about, published conference papers can often be considered serious contributions to the literature, and will get cited.
This is really field dependent. In CS, aside from maybe graphics and computer vision conferences are considered much less serious and prestigious than journals.
Edit: Certainly things from conferences will get cited if they’re directly relevant to your new work, but getting into a good journal is all around “better”.
Of course it is.
The distinction between “conference” and “journal” is not the issue. The issue is poor quality publications and higher quality ones. Some of the best papers in CS were only published in top conferences and never submitted to journals. Most journal papers are rarely read or cited.
There are too many phony conferences popping up. Some with names similar to respectable ones. The phony paper demos just show the extant of that problem.
Protecting the reputation of all publications, conference or journal, is important.
There are conferences and there are conferences. I’ve founded a few, and been program chair for some others, all IEEE. There are no IEEE regulations about how peer review for conferences or workshops are done. Some use multiple reviewers for full papers, some use the program committee for abstracts.
Those in the field know which is which.
I submit papers from the conference I’m most involved with to IEEE for inclusion in its online library, and I assure you that if you use their software correctly to define the contents no one there looks at the papers.
This goes to show that you should no more trust papers from China than you do toothpaste.
I suspect that any discipline has small conferences desperate for papers which might take anything without looking too hard - especially of the conference organizers are clueless and are doing it for clout.
Loved the paper. Many of the authors are famous computer scientists, including Alan Perlis who died 9 years before his supposed paper. Also loved the network of PDP-11s.
That sentence almost makes sense - I can see dogfooding as using what you propose is good for work for work. My company moved all its accounting systems to the computers we made, and the CEO said this was an example of eating our own dogfood.
I wonder how they generated the charts. Did they pick some at random, or did they randomly fill in boxes. I was hoping for a flow chart box with no inputs or something.
Computer science is different in that conference proceedings are seriously refereed and, in many cases, constitute the bulk of a researcher’s publications. The situation in math is quite different in that papers published in conference proceedings are hardly counted and only journal publications matter (for tenure, promotion, research grants). I had them generate their generic math paper and it is obvious gibberish from the merest glance. It was grammatical (more or less) and used words that mathematicians might use, but not in any sensible combination.
Unfortunately, a lot of new journals are springing up almost daily. You pay the journal to publish your paper and they distribute it for free online. You are expected to produce a finished pdf file and they “publish” it. A week does not go by without my getting an ad for one of them. One of the things they offer is quick publication, which is a sure sign they are not refereed, at least not seriously.
Obligatory XKCD reference.
As I said, there are conferences and there are conferences. I put together a session for the ACM Computing Science Conference in the early 1980s and if I don’t recall getting an reviewers at all. Getting the papers was hard enough. The conference was basically for recruiting new PhDs.
Based on the concern of the academics on our Steering Committee, tenure committees evaluate conferences in no small part by paper acceptance rates - the lower the better. I bet the accept rate for the conferences who took these papers is close to, if not over, 100%.
BTW, I’ve seen submissions by non-English speakers which are not much less coherent than the funny CS paper linked to. (I also noticed that the authors listed in the text as being associated with a reference weren’t.) Hmm - maybe some people took this paper as a template. :dubious:
According to the Nature article, most of these papers were published in the proceedings of conferences held in China. I don’t know whether or not conference papers are normally peer-reviewed in China, or if that was the case for these specific conferences. The article cites the website of one of the conferences as saying that papers are “reviewed for merits and contents”, but that doesn’t necessarily mean they were peer-reviewed.
I’m not sure how legitimate any of these Chinese conferences were, but I can believe that even a well-intentioned reviewer who didn’t speak English that well might not have looked too carefully at an English language paper that had a respectable scholar’s name on it. The article indicates that some papers used the names of actual scholars, but that the real people may not have known anything about this.