A prime number of at least 2 million decimal digits will be discovered, and confirmed by reputable third party, on or before 31 december 1999.
The claim will be considered confirmed if reported in the popular press such as Scientific American, Science, Nature, or any of the well known mathematical journals.
I plan to judge this claim primarily on the basis of intent if there is a conflict between the strict wording and the perceived intent. The reason for this is that I was in close contact with the author during the planning phase, and have a good idea of what he had in mind.
If a credible claim is made of a 2,000,000 digit prime before 2000.1.1, I will wait until judgment date if necessary for confirmation to be published; if it's forthcoming, the claim closes early. Probabilistic tests will not satisfy this claim: primeness must be proven. If a credible claim is made between 2000.1.1 and the claim's judgment date that such a prime was discovered before 2000.1.1, I may defer judgment for up to another month to allow confirmation. Credible claims that could trigger a deferred judgement would include claims by GIMPS or by a number theorist whose name I recognize; but in no case will a prime discovered between 2000.1.1 and the judgment date qualify. Random claims by unknowns (e.g. FX punters) will not trigger a deferred judgment.
I intend to take a broad view of what constitutes a mathematical journal: the thrust of the claim is that the prime be discovered, not the precise manner of its reporting. If it's certain by the judgment date that such a prime was discovered in the correct time frame I will declare it a YES even if none of the indicated sources has had time to get it into print.
Caveat Mercantor: The Great Internet Mersenne Prime Search (GIMPS) has announced at http://www.mersenne.org/6972593.htm that the 38th known Mersenne Prime was discovered and verified in June 1999, and that it is over 2 million digits long. Publication should follow shortly.