The Speculist is musing on a question posed by Chris Anderson, on whether petabyte computation obviates the need for theoretical development.
Now, I’m, as my twin points out, not an analyst, but I DO a lot of work in theory (often reaching wayyy out into heresyland), in areas where such computational ability is of little use. Though, it must be said, most of my hypotheses actually are falsifiable and/or rest upon questions where the data is empirical. So, subtracting out field-selection bias, what do you guys think?
On the one hand, it’s a ridiculous question — but ridiculous questions often yield useful fruit.
(Speculist guys, still lovin’ your page, still having trouble commenting).
JimDesu
/ June 26, 2008– Petabytes allow us to say: “Correlation is enough.”
WTF? Correlation is enough?! Yeah, and objects fall because they’re seeking their own ordained places in the heavenly spheres. Oh, and forget about elliptical planetary orbits, let’s just have tables upon tables upon tables of corrections to use data instead….
If you want to drive your society forward while looking in the rearview mirror, fine, but not me.
Andrew Reyna
/ June 26, 2008This guy is starting to sound like Max Cohen from Pi. Compute the future, existence is just a string of algorithmic patterns.
RandomDude
/ June 27, 2008http://science.slashdot.org/article.pl?sid=08/06/25/146250&tid=217
Alex
/ June 27, 2008Petabytess of data just allow you to look at several correlations and make better grander hypotheses to test and convert into theory.
Data is just data – it means absolutely nothing unless you can make sense of it, and while you can correlate trends, this is just data grouping. Looking at the groups and saying “Aha! That’s why that happens, now let me predict something and see if I’m right.” …that’s Theory, and correlation does not replace it.