Submit late. This is the basic rule in having your abstract rejected. Don't even start writing it until the deadline for submission is long past. Keep the program committee informed of your progress. ``Seems to be a little hole in the proof somewhere.'' ``Don't sit on the edge of your chairs.'' ``Almost ready.'' ``It's a-comin'.'' ``Any minute now.'' Everyone on the committee is sure to remember your name when your abstract finally arrives.
Submit incorrectly. The device of sending abstracts to the local arrangements chairman is overused. Try something fresher. Send your abstract to last year's program chairman. Send it to this year's in care of the school where he did his undergraduate work or, better yet, to the school that turned him down for tenure. Send it to someone whose name sounds a little like his. Under any circumstances, be sure to send it postage due.
Grossly exceed the maximum length requirements. Most extended abstracts should be eight to twelve pages long, or between 1,500 and 3,500 words. Your aim, then, should be for at least 10,000 words. (Read symbols aloud to count how many words they are -- don't count characters.) There are several interesting variations on this ploy.
The Godzilla: Submit a seventy-page paper with instructions to - the program committee to read the first twelve pages. Be sure page 12 ends mid-section, mid-paragraph, mid-sentence.
The Monster from the Black Lagoon: Submit a twelve-page abstract with thirty pages of appendices. Be sure there is no way anyone can understand the body of the abstract without reading all of the appendices. By-far the easiest way to accomplish this is to introduce your own utterly idiosyncratic notation. 1+1 = 2, for instance, will in your notation be written
b b| < 1 2 1*. z| > ^
The King Kong: Submit an eight-page abstract of 20,000 words. You may need special typographic equipment for this one, but don't worry; it exists. With an IBM composer, six-point type, no margins, and no displays, you can write 20,000 words on the head of a pin.
You might think that the opposite strategy would work equally well --- submitting an abstract that falls far short of the minimum requirement. Not so. Look at it from the program committee's point of view. They must read a hundred or more abstracts in the midst of their other duties. Mere brevity after all those monster abstracts is going to look good to them.
Vacuity, however, is an excellent technique, and it may be allied
with the shortness strategy for a Run Spot Run abstract. Here
is a good example:
We worked in complexity. We proved some theorems.
We proved some big theorems and some little theorems.
Some proofs were big, some were small. We tried
to match up the proofs with the theorems, but we
couldn't always do it. Then we were sleepy and went
to bed. Good night.
This is a good example but not a perfect one. Substitute ``computer
science'' for ``complexity''" and you will see how much room for
improvement there is in any abstract, no matter how vapid it may
seem at first glance. Only by constant, careful revision can
you insure the rejection of every single abstract you prepare.
A new tactic we would like to commend is the Grocery List. For this you must give at least forty theorems. The North American record stands at 97, but there are allegations that the author was under the influence of chemical stimulants and the judges are currently reserving the title. We confidently expect to see in next year's competition exciting new combinations of the Grocery List with the Run Spot Run and other devices.
Give no motivation. Present your results in a vacuum. Strip your ideas of any hint they might offer as to their origin, direction, or relevance. Say nothing about practical applications unless you are submitting to a theory conference, in which case you should be sure to call them ``pragmatics''.
Be sure also to give no background. Even the novice will know enough to leave off all acknowledgements and references. But the master will go further. He will give the appearance of citation without any substance. He will enclose a reference 1ist on which every item is submitted, in preparation, or a private communication. He will call obscure results by pet names he has invented himself. And he will describe as ``well known'' results published only in Old Serbian --- preferably false ones.
Prove trivial results in exhaustive detail, breaking your proofs into as many lemmas as ou can and disrupting the line of reasoning with notes, remarks, and asides. On the other hand, assert difficult proofs. Assert them badly, with a sneer, if you can manage it. The judicious typographical error in the statement of your theorem adds a note of drollery to this device.
Never under any circumstances provide a cogent verbal sketch of a proof that stresses its provocative turns while leaving the obvious unstated. Never. This alone may get an abstract accepted even if you have faithfully followed all the rest of our advice.
As for your language, it should be pompous, impersonal, drab, and bleary. Work hard at your grammar. There is no excuse for agreement between subject and predicate in any sentence of more than ten words. Indefinite referents combined with false parallels will leave the unwary program committee member clutching his head and wheeling about the room in confusion. Any temptation he had to accept your abstract will disappear instantly.
Mind the appearance of your paper, too. An ancient, faded grey typewriter ribbon is good, but why not try a fading red one for that extra spark of individuality? Sixteen-pound paper seems flimsy enough until you realize that you can find twelve and even less. Why not type on the back of used second sheets and enclose a sanctimonious note about recycling paper products? Why not submit a handwritten manuscript --- written in pencil? Why not have your four-year-old type your paper? And what ever became of the old-fashioned muddy footprint, the coffee ring, and the grease smear? You cannot give too much attention to these small details. They can make the difference between an abstract that is marginally, reluctantly acceptable and one that will be firmly rejected year after year.
(You should understand that once you have a soundly accept-proof manuscript you should resubmit it every year. You will become part of the mythology of your field. As program committee succeeds program committee, the question will be asked, ``Did you get Old Whosit's paper again? What's he calling it this year?'')
A final word: If after adopting all these strategies and developing several of your own you have a paper accepted anyway, do not despair. Do not take it personally. The program committee has certain quotas it must fill. A certain number of papers must be accepted regardless of merit. Your friends and colleagues will understand, and no one will hold it against you. Just don't let it happen too often.