All right, I’ll take a stab at a couple of these, starting with (k): So prior to recombination, the universe was almost completely ionized. That is, the temperature was high enough that if an electron happened to bind to a proton or heavier nucleus, it would almost immediately get knocked back off. Now, photons interact electromagnetically, and since almost everything they were travelling through prior to recombination was ionized, they were constantly being scattered.
Then, over a relatively short period of time, the temperature fell enough for neutral atoms to form. The charges are all still there, but an atom looks neutral until a photon gets close enough to see the separation between the electron(s) and the nucleus, which is awfully close. Hence the overwhelming majority of the photons that existed at the time of recombination have been “free-streaming” since then, and some of them land on our detectors. There are many things that can slightly alter the CMB spectrum, of course, such as scattering off of hot gases in galaxy clusters, but these make only minor modifications to the spectrum.
Now, how do they measure cosmic parameters (and hence things like the whether the universe will expand forever, its constituents, and so forth)? This information comes from looking at the CMB power spectrum, an example of which you can see here. Essentially this is a measure of how those fluctuations you see on the maps they show are correlated (note that the maps are actually showing deviations from the mean CMB temperature). The spherical harmonics on the x-axis correspond inversely to angular scales; that is, a smaller l is a larger angular separation on the sky.
Where does this kind of structure come from? Structure formation occurs on a range of scales in the universe. The CMB is a picture of the universe at 380,000 years old. Since the big bang, the matter in the universe had been interacting gravitationally, with regions collapsing if the gravitational force was strong enough to overcome thermal-pressure forces. On some scales there had been enough time by recombination for a cloud of matter to contract to its maximum (given its makeup, temperature, etc) density. On some scales matter had not had time to do that. On some scales it had had time to contract, then rebound. And then perhaps contract and rebound again, and so on. The power spectrum shows what the level of correlation, or structure formation, on different angular scales was at the time of recombination.
Now, to fit a curve to the data points requires that one have a cosmological model built from general relativity and quantum mechanics, describing the structure of spacetime, the matter and energy constituents of the universe (baryonic matter, dark matter, dark energy, etc.), the rate of expansion and how quickly that rate is changing, and so on. By fitting a model with those parameters to the data, the WMAP team has come up with the various numbers given for the universe’s makeup, the Hubble parameter, the age of the universe, its geometry and fate, and so on. If you look in the technical papers section, the “Angular Power Spectrum” paper has previous CMB measurements plotted on page 37, and the WMAP results page 38 onward; and the “Determination of Comological Parameters” paper gives their results for parameters.
I don’t know where the result about the first stars comes from–they say it has to do with measurements of CMB polarization, but maybe someone here can explain that.