The setting is 2036 to 2038. A 140-km long linear collider ("Super Large Hadron Collider") has been installed at one of the L5 points in earth orbit. Some unknown technology must have been discovered since the accelerator is able to probe the Planck regime. So once the accelerator goes operational and creates big-bang level energies, something appears to change at a very fundamental level. Indeed, since there is good reason to believe that causality has no meaning at such high energies, the accelerator spawns off a new universe. The new universe has all of the memories and artifacts of the old one but starting from its own moment of creation, a new physical reality has also started seeping in. In particular, new rules of mathematics have come into play, causing inexplicable phenomena around the world over the next 18 months...the seepage must be slow (all if this is quite understated in the story, unlike the breathless story-telling of a Nat Schachner or Russell Fearn).

A computer programming genius finds out that the principle of complementarity - existence of inverse operations in arithmetic - is no longer true. Carrying out calculations in a system of equations in two different orders give different answers. These "inconsistencies in complex systems of equations vary with time and geographical location." As he muses: "There's a saying that the book of nature is written in the language of mathematics. [...But] perhaps mathematics is not a realm of ideal objects after all, but firmly yoked to material reality, so that when that changes, mathematics changes, too. Something must have happened to change the structure of reality. [...] Soon mathematics as we know it will no longer exist"

I really quite liked the story for its basic idea and exposition, though the characters and plot line could have been much better. His use of the failure of the axiom of identity ("x is suddenly not x") makes all of the new mathematics inconsistent, which seems quite drastic in the face of the continued existence of the new universe... . Nonetheless, highly recommended.