RealTechNews.com - Independent Tech

Sunday, October 10, 2004

New Tech, Old (and bad) application.
It isn't until you get to the end of a rather interesting story detailing a new technique for creating a registers in quantum computer environments (using atoms rather than trapped ions) that you hit this sentence:

"One potentially important application for a quantum computer would be factoring the product of two large prime numbers. This would make it instantly possible to crack the cryptographic codes used to secure many computer communications."

Well, the scientists involved in this thing think it will be at least two more years before they can get two registers to interact with each other consistently, but, in the meantime, the question here is: Are we going through the trouble and expense of developing quantum computing just so someone can hack into somebody else's system? I can only wonder, with all of the other wonderful things that might be possible using quantum computing, why the newscientist.com writer went to that issue?

-
# Permalink Posted at 1:40 PM 6 comments Email this Link

Comments on this Item:

That one always get cited. I think it is because it is fairly easy to state and it has a nice alarming ring to it.

# posted by Davre 8:17 AM

It is simply that the writer is aware (or believes) that everyone understands that this computing task takes a long time with standard processing hardware. What everyone who views this with alarm forgets, is that should "quantum computers" become affordable for any other than large governments, they will soon trickle down to the masses, so those that want/need to secure their own communications will be able to use ridiculously huge primes for you security, that will again secure your communication.

# posted by Anonymous 8:29 AM

Yes, this is alarming. However, there are already quantum algorithms for encryption that are very secure in addition to those algorithms that decrypt large prime-based encryption like RSA easily. All this means is that RSA won't be used anymore and these quantum encryption algorithms will replace it.

# posted by Anonymous 8:40 AM

Actually, at the moment no one would really know what to do with a quantum computer if we built one. There are only two things we know a quantum computer could do faster than current computers: factorization (or more precisely, solving a whole class of problems known as Abelian hidden subgroup problems, of which factorization is just one example), and solving NP problems slightly faster (so-called database search, though there are issues with using it to search an actual database).

(There are also a few very new ideas based on quantum random walks. And a quantum computer might be useful for simulating quantum systems which if you're optimistic might help understand high-temperature superconductivity and the like, though that hasn't been proven rigorously.)

The "database search" algorithm doesn't speed things up as spectacularly as factorization, which is maybe why it isn't mentioned as often. And "Abelian hidden subgroup" might be more precise and more significant, but it makes a lot less sense to most people than "factoring"!

Why go to the trouble and expense? There are probably many other interesting things one could do with a quantum computer - it just needs someone clever enough to think of them! Since it's going to take a while (decades) before anyone builds one, there's plently of time to get thinking.

More generally, whether you call it quantum computing or quantum something else, the experiments are really excercises in controlling quantum systems more precisely than has ever been done before. It's fundamental research, and history tells us that applications usually come from completely unexpected quarters. (Remember the laser, another quantum system, which a certain famous physicist called "a solution looking for a problem"...?)

Toby Cubitt

# posted by Anonymous 8:56 AM

Wait, what are the odds here... The previous poster's last name was Cubitt. as in Qbit, as in quantum bit, making a comment about quantum computing with 5 qbits.

And his first name, Toby... That my first name as well, and not at all common.

But anyhow, on topic, it seems that quantum computers combined with neural nets would be efficient at computing anything that can be done currently through repetition. i.e. video encoding, interplanetary ballistic solutions, seti@home, etc, where there may be (inset large number here) of solutions, but only one of which is optimal. Maybe. We shall see..

Toby Hackstock

# posted by Anonymous 9:42 AM

And if Microsoft created an operating system for a quantum computer, would it be called "Black Hole?"

# posted by Bill 12:04 PM

Post a Comment