For years, researchers had presumed that Langmuir monolayers of small C(n)F(2n+1)C(m)H(2m+1) (FnHm) diblock molecules (such as F8H16) consisted of continuous, featureless films. Recently we have discovered that they instead form ordered arrays of unusually large (~30-60 nm), discrete self-assembled surface domains or hemimicelles both at the surface of water and on solid substrates. These surface micelles differ in several essential ways from all previously reported or predicted molecular surface aggregates. They self-assemble spontaneously, even at zero surface pressure, depending solely on a critical surface concentration. They are very large (~100 times the length of the diblock) and involve thousands of molecules (orders of magnitude more than classical micelles). At the same time, the surface micelles are highly monodisperse and self-organize in close-packed hexagonal patterns (two-dimensional crystals). Their size is essentially independent from pressure, and they do not coalesce and are unexpectedly sturdy for soft matter (persisting even beyond surface film collapse). We and other researchers have observed large surface micelles for numerous diblocks, using Langmuir-Blodgett (LB) transfer, spin-coating and dip-coating techniques, or expulsion from mixed monolayers, and on diverse supports, establishing that hemimicelle formation and ordering are intrinsic properties of (perfluoroalkyl)alkanes. Notably, they involve "incomplete" surfactants with limited amphiphilic character, which further illustrates the outstanding capacity for perfluoroalkyl chains to promote self-assembly and interfacial film structuring. Using X-ray reflectivity, we determined a perfluoroalkyl-chain-up orientation. Theoretical investigations assigned self-assembly and hemimicelle stability to electrostatic dipole-dipole interactions at the interface between Fn- and Hm-sublayers. Grazing-incidence small-angle X-ray scattering (GISAXS) data collected directly on the surface of water unambiguously demonstrated the presence of surface micelles in monolayers of diblocks prior to LB transfer for atomic force microscopy imaging. We characterized an almost perfect two-dimensional crystal, with 12 assignable diffraction peaks, which established that self-assembly and regular nanopatterning were not caused by transfer or induced by the solid support. These experiments also provide the first direct identification of surface micelles on water, and the first identification of such large-size domains using GISAXS. Revisiting Langmuir film compression behavior after we realized that it actually was a compression of nanometric objects led to further unanticipated observations. These films could be compressed far beyond the documented film "collapse", eventually leading to the buildup of two superimposed, less-organized bilayers of diblocks on top of the initially formed monolayer of hemimicelles. Remarkably, the latter withstood the final, irreversible collapse of the composite films. "Gemini" tetrablocks, di(FnHm), with two Fn-chains and two Hm-chains, provided two superposed layers of discrete micelles, apparently the first example of thin films made of stacked discrete self-assembled nanoobjects. Decoration of solid surfaces with domains of predetermined size of these small "nonpolar" molecules is straightforward. Initial examples of applications include deposition of metal dots and catalytic oxidation of CO, and nanopatterning of SiO(2) films.