"Instant" Scientific Article Exposes a Flaw in the Peer Review System

Forrest M. Mims III

Publishing papers in the peer-reviewed literature is by far the best way for a citizen scientist to earn credibility. Once you have published some serious papers, you will have entered the world of professional science, even if you lack the usual academic credentials.

To beef up my publication list, I have just generated a new computer science paper in record time. Here is the abstract:

Investigating Symmetric Encryption and Virtual Machines Using Amt

By I. R. A. McFakePaper

Abstract

Many experts would agree that, had it not been for the Ethernet, the emulation of architecture might never have occurred. It at first glance seems counterintuitive but usually conflicts with the need to provide simulated annealing to cyberinformaticians. In this work, we disconfirm the confusing unification of IPv4 and flip-flop gates. In our research, we prove that operating systems and replication are always incompatible.

Most papers require months or even years of effort. This paper required only about eight seconds, the time necessary to enter a fake name into the author space and to click a "Generate" button on the web site that helped prepare the paper. Actually, the web site prepared everything but the fake name. It automatically selected a title, wrote the randomized text and inserted the randomly generated plots ands charts that are arranged nearby. It even generated a perfectly organized and formatted list of randomly generated fake references.

The paper looks so good that I decided to write another one. This required even less time, since my name was already in the author box. Before my fingertip left the mouse button, a completely new paper was on the screen. It's title is "Refining Superpages Using Stochastic Symmetries." It's a super paper.

If two, why not three? My third paper this morning is entitled, "Visualizing RAID and Byzantine Fault Tolerance."

While my three new papers are totally fake, they all look great. I would be proud to send a scientific conference committee papers as neatly formatted as these.

For the past several months I have used every spare moment to work on three real papers. That's why these three fake ones have significantly improved my morale and outlook. There was no need to do endless experiments and spend countless hours processing and analyzing reams of data. It wasn't necessary to make carefully planned graphs. Nor was it necessary to spend hours developing a list of credible references and then typing them all in the same format. It wasn't even necessary to think of an idea, create a clever title or write a carefully crafted abstract. A spell check wasn't even required.

Why go to all this trouble when my randomly generated instant papers have sufficient quality to be accepted by at least one computer science conference? That was recently demonstrated when the MIT graduate students who developed the software that wrote my three fake papers recently sent two of their fake papers to the World Multi-Conference on Systemics, Cybernetics and Informatics (WMSCI). Much to their surprise, one of their fake papers was accepted (Jeremy Stribling, Daniel Aguayo and Maxwell Krohn, Rooter: A Methodology for the Typical Unification of Access Points and Redundancy).

You can find out all the details about this at SCIgen - An Automatic CS Paper Generator. You can even generate your own fake papers from this web site. Don't be concerned if you have never written a scientific paper. I just wrote a fourth fake paper using the name of our neighbor's dog, who has never written a scientific paper or anything else that I know about. The title of Buddy's paper is "Comparing 802.11 Mesh Networks and Agents Using Furze."

Buddy's abstract reads, "Unified pseudorandom information have led to many natural advances, including erasure coding and agents. Here, we argue the investigation of XML. We examine how SMPs can be applied to the simulation of gigabit switches."

This sound so intriguing it must be real, so I'm thinking about either joining Buddy as a co-author or switching his paper for one of the three I wrote earlier. After all, Buddy will never know the difference. And, apparently, neither will at least one computer science conference.

For more about the adventures of the MIT students and their fake paper, see "MIT Students Trick a Computer Science Conference" in this issue of The Citizen Scientist.