Papers with unorthodox writing style

I'm not sure if this is the right forum for this question, in any case probably CW is appropriate?

I've been looking around the mathblogosphere for the past few weeks and ran into mathgen. It's pretty amusing, to be sure, but point 6 in the Why? section has set me to serious thinking. For those opposed to clicking links, mathgen is a random generator of math papers, and the creator give several justifications for the creation, the relevant one being:

I think this project says something about the very small and stylized subset of English used in mathematical writing. This program only knows a handful of sentence templates, and yet I think its writing style is [typical.] I think we could stand to pay more attention to our writing styles, instead of unthinkingly relying on stock phrases.

With this in mind, have any of you encountered reputable, "research-tier" papers that have a writing style dramatically or at least distinctly different from the one that seems to dominate so much of this kind of mathematical writing? I'm not really looking for expository writings, although I imagine that what I am looking for will have a similar feel to it. So I think what I'm going to mean by research-tier (for now) is simply that it proves something new and at least mildly significant.

Links, especially free ones, are appreciated.

Solution 1:

One piece that comes to mind is Gromov's "Metric structures for Riemannian and Non-riemannian spaces." Everything from the numbering - a Gromov hallmark - to some of the colorful yet sometimes remarkably illuminating language - a wonderful example is 1.25.1/2, "If one feels disgusted by the spineless flexibility of arc-wise isometric maps, ..." - is rather unique in the literature. (And still, Gromov is far more digestible in print than when speaking.)

Solution 2:

This is moderately common in Theoretical Computer Science. A lot of problems and algorithms in that field are posed and discussed in terms of mini stories, even in technical papers. In his celebrated paper introducing what is known as the "Arthur-Merlin Protocols", Laszlo Babai writes as if telling a fable about Camelot.

Leslie Lamport, the founder of the field of distributed systems, wrote a 33 page paper describing a protocol with the entire thing wrapped in a conceit about a fictional greek society's parliament deciding on laws to pass, which itself involves a nested conceit about a bunch of Greek priests agreeing on a decree. Three years later published a paper outlining the same protocol but in a more normal tone and it was a third the length. He even makes a quip about the fact that the original paper was Greek to many readers and was therefore inaccessible.