Two scientists develop a mathematical method of literary analysis based on the use of an "author function". The zeroes of this function (called Nullstellen in German, as in Hilbert's famous Nullstellensatz) uniquely identify the author. They attempt to create a text having no zeroes, an "authorless story", with literally disastrous results.
To give you an idea of the sort of mathematical terminology in the story, here is of one of the scientists describing her PhD thesis:
(quoted from Nullstellen)
The evaluation of 2nd order Legendre functions. I develop an evaluation
method applying the trapezoid rule to Heine's integral representation,
including error analysis, with some numeric results. Nothing fancy,
quadratures methods, I omit Gaussian or ClenshawCurtis quadratures.
Not very precise up to now, and not much of practical use.

The story was first published in SPEX (a magazine that Dath edited) in October 1999 and reprinted in his collection Heute keine Konferenz in 2007.
Contributed by
Hauke Reddmann
[Dath's writing] at its best can give you ontological goosebumps like otherwise only
Philip K Dick. Even better, the text had to fit in an illu. Like all geniuses
he likes to hear himself blabber (I speak from personal experience :) and
the short form becomes him.

Thanks to Hauke Reddmann for informing me of, summarizing and offering a translated excerpt from this story. 