something wrong here?
 

New Scientist recently published an article about quantum computing, entitled "[url=http://www.huliq.com/34160/qubits-poised-to-reveal-our-secrets]Qubits poised to reveal our secrets[/url]" (I didn't use the New Scientist official site since it requires a subscription).

However, somewhere in the article, it says:

[QUOTE]For instance, to find the prime factors of a 10-digit public key, approximately 100,000 calculations are needed; for a 50-digit number about 10 trillion trillion are required. IBM’s Blue Gene supercomputer would take a fraction of a second to crack a 10-digit key, [B]but about 100 years for a 50-digit key[/B]. And keys are now much longer than 50 digits. In 1994, mathematician Peter Shor at Bell Labs in New Jersey developed a routine that radically reduces the time required to make those calculations. There was just one rather large catch: it could only run on a computer that exploits quantum mechanics.[/QUOTE]

Huh? I think there must have been a mistake or something, as people have factored 200-digit numbers.

It states that Blue Gene would take that long, but I assume that they are referring to a brute force method.

[quote=ixfd64;114418]Huh? I think there must have been a mistake or something, as people have factored 200-digit numbers.[/quote]
I was going to write about how the key they're talking about likely has many more than 10 possible combinations per digit, but then I realized that even a 256 digit key with 50 possible combinations is in the 400-digit range, and wouldn't be so hard to factor, not 100 years of Blue Gene at least, even if semiprime.
So...I think I agree with you that there is a mistake there.

EDIT: I just had my (quite old and slow) computer factor a 51 digit semiprime number, and it did it in 1 minute 34 seconds, with me doing other things on the computer. So, in any case, it's definitely not talking about a semiprime that's about 10^50.