When Einstein developed the general theory of relativity he realised that it meant a static universe was impossible - gravity would force such a universe to start shrinking. Hence he added in the 'cosmological constant', an anti-gravity term which he intended to allow the universe to remain static. Hence from the beginning, some cosmological models based on general relativity included a nonzero cosmological constant, for instance the de Sitter universe. When Hubble found that the universe isn't static after all, Einstein called the 'cosmological constant' his greatest blunder. But somehow it wouldn't go away.

One problem in the middle of the 20th century was that the Hubble expansion implied the age of the universe was less than that of the Earth. Adding a cosmological constant allowed an older universe, and so many cosmological models of the time included such a term. As the Hubble constant was measured more accurately this was found to be unnecessary. However in the 1990's evidence evidence from distant supernovae implied that the expansion of the universe was accelerating. It is thought that the universe is full of dark energy which is causing this acceleration, and so the cosmological constant is back to stay.

Event horizons are usually considered as belonging to black holes - once anything crosses such a horizon there is no return. However, an event horizon is also predicted in cosmological models with a positive cosmological constant - distant matter can be accelerated away and so become permanently out of contact with us. There are some differences between the two types of horizons - for a black hole the horizons is a one way barrier in that you can cross it but can't return. The cosmological horizon on the other hand works symmetrically - if B is out of the reach of A then A is out of reach of B. In many other aspects however it is similar to the event horizon of a black hole.

In the 1970's Stephen Hawking published a paper which amazed everyone as it predicted that black holes weren't actually black - they would have a non-zero temperature and so emit radiation. Less well known is a calculation made a couple of years later showing that a cosmological events horizon will also have a positive temperature and produce radiation. At the time it was thought that there was zero cosmological constant and so no cosmological event horizon, and the paper was of academic interest only. However now we know differently - current cosmological data implies a horizon.

Black holes are weird objects, but they are becoming better understood - or at least we are getting used to them. The way that time seemed to stop at the horizon is seen to be merely due to coordinate choice. Theories of quantum gravity are judged on how well they deal with the singularity at the centre. The 'information paradox' has been around for a couple of decades, but seems to be being resolved. However, the cosmological event horizon opens up another can of worms. At least there's no singularity, but the problem of time dilation at the horizon looks worse - we don't have anything corresponding to an infalling observer who doesn't see anything strange about a black hole horizon. Furthermore, if there is an information paradox here then is it harder to see how it can be resolved - information which has crossed your horizon it is seemingly lost to you forever. And while a black hole will eventually evaporate away, Hawking radiation means that the cosmological horizon comes closer to you, making any problems worse. (More strictly, the event horizon will approach the observer if the detector absorbs radiation, thus increasing its mass, and there is a natural limit to this. In the words of Hawking and Gibbons He has to be careful that he does not absorb so much radiation that his particle detector undergoes gravitational collapse to produce a black hole. Something for experimenters to bear in mind?)

However, the cosmological horizon isn't all bad news. In a sense it defines a boundary to our universe, which means that that instead of having to deal with all of the philosophical problems of an infinite universe we can confine our theories to a finite volume. For instance in discussions of quantum theory there is the question of what is meant by the wavefunction of the universe, and its much easier to make sense of this if the universe is finite. Likewise the entropy of the universe makes more sense if there is some sort of boundary, especially a boundary with thermodynamic properties. And when the problem of infinity is out of the way it is possible to relate the entropy of the universe to event happening within it such as the origin of life and its eventual fate. The cosmological event horizon hasn't shown up very much in works of popular science until now, but I feel it is likely to become much more widely known in the near future.