Large Hadron Collider

The LHC, which goes online tomorow, is going to be capable of reproducing the conditions shortly after the big bang. Media attention has been brought to the fact that it also has the chance of creating "mini" black holes, in which case, some have speculated that these high-energy conditions could in fact backfire, blowing the Earth to smithereens.

Stephen Hawking disagrees.

Anyways, while reading an article about it, and the grid compute farm that will process its data, I ran across this Gem of prose:

When the experiments get running at the LHC, the four great "eyes" of the machine start observing collisions, they will generate 15 million gigabytes of data every year, that is equivalent to one thousand times the information printed in the form of books annually.

"If you put them on CDs and stacked them up, that stack would be more than 12 miles (20 kilometers) tall." said Dr Bob Jones, Director of the EGEE, Enabling grids for e-science project, which is co-funded by the European Commission.

Or, in terms of iPod data, the annual output of the atom smasher is equivalent to a song running for 24,000 years.

OK. "15 million gigabytes" of data I get. I understand that because my laptop has a 160GB. My backup drive has 500GB. So 15 million GB pretty much blows my mind, but I get it.

So why, then, must this writer give us not one, but THREE, interpretations of what that amount of data means when most people reading about the Large Hadron Collider already get it? 1000 times the amount of information published in books annually?? How am I supposed to comprehend that? How do I have any reference in my head about how much information is published on an annual basis in the form of printed material?

How about the great visual of stacking CD's on top of each other? I suppose that 12 miles of CD's stacked on themselves is such an impressive visual that it gives me a great appreciation for the amount of data that truly is. Unfortunately, it doesn't. Fail.

And the true stroke of genius, the example that should win this guy a Pulitzer, the boiling it down the "iPod data." Take a moment and let the absurdity of that sink in. Now let me paraphrase it for you -- "If the annual amount of data from the LHC were TURNED INTO A SONG, it would play for 24,000 years." Thank you for that. This has actually done more to CONFUSE me about the magnitude of the data than anything.

Confusing examples aside, 15 million GB of data annually is an extraordinary amount of data to be processed, and the technology to support that (the 80,000 node grid farm) is in itself equally amazing.