Deepak

I must admit that I was very underwhelmed by the article. Chris is a little out of his league isn't he? - Deepak via Bookmarklet

We were crunching the genome with huge clusters a decade ago. As we get more information, our models get better. To make more sense of this information our theory has to get better to as Lee Hood always says. Ah ... I am bugged. All Hypy with limited fundamental understanding - Deepak

Yeah, pretty naive stuff. - Neil Saunders

I'd like to see someone try to sell 'correlation is enough' to the FDA or equivalent - Cameron Neylon

Cameron: lol In journalistic terms the essay is good, but it's science where being only trend sensitive is not enough. And Anderson has not much clue on what's going on in biology, but he is probably right on what's going on in physics, which we don't really know. - Attila Csordas

Am reading the paper version of Wired at this very moment! The cover uses the headline "The End of Science." Now that caught my attention as I pulled it out of the mailbox! But yes, underwhelming. I don't think scientists are going to stop asking "why?" anytime soon, the way the article makes it sound. - Shirley Wu

like I said last time this came through - it's a tad grandiose. Did he just learn about data mining recently? - Mr. Gunn

Ah - plane travel tomorrow which is usually when I buy a copy- mind you probably be last months on this side of the pond anyway :) - Cameron Neylon

A good response: http://bactra.org/weblog/581.h... - Bill Hooker

I agree, Bill. Makes one think of the many propositions to replace wet lab experiments with in silico substitutions. You need both. - Heather

Chris still is onto something. In industry you can create value from correlations and "hunches" based on data. Even medical trials are just data and correlations. And in oder to save lives you don't really need to know why, just what. Science is a method, never forget that. Not a religion. And its limits are there, as well as its advantages. Other methods might be more efficient at doing something else, not? - dekay

Actually, that used to be the case. The focus these days is on the why, on the mechanisms. It's why VIOXX happens, cause you don't understand the why. Yes, you need statistical separation, you always have, even when you had limited data. More data only improves your statistics. - Deepak

For example, docking studies are increasingly trying to add more "physical" measures rather than rule based implicit, this is what the data tells us measures. Science is much too noisy, and the kinds of false positives you would get with Google are NOT acceptable. Is the scientific methods dogma. Can it change, yes, but not for the reasons that Chris is talking about - Deepak

Even Valleywag has an opinion on this, they are scientists too, and have a lab at the back: http://valleywag.com/5019748/w... "The problem here is that if we stop asking the question "why?" then we are basically making for the foundations of faith. You can always make statistics say nearly anything you want, it simply depends on the assumptions you make when you analyze and present them" - Attila Csordas

When Valleywag is right, it's a sad day - Deepak

Attila - no, Anderson is definitely not right about physics, either. It's a strange article. Data mining isn't replacing the old scientific method. It's augmenting it. - Michael Nielsen

Agree with Michael. The writer seems to think that somewhere in a big pile of data lies "everything", waiting to be extracted using algorithms. That's just not how science works. It's the questions that we ask and the interpretations that we devise that make it science, not the raw information. - Neil Saunders

I created a slide at my previous job, which had all the data sources (expression, proteomics, genotyping, histopath, name it) that exist related to a study on the edges, with the emphasis on there being a ton of heterogenous data, and in the middle a big "?". The point. It's all about the question, NOT the technology or data type - Deepak

Chris' piece is provacative ... but the Ars Tecnica piece touches on some key points - was /.-ed a bit ago, so my apologies if this is old news: http://arstechnica.com/news.ar... - Kaitlin Thaney

Kaitlin, that is a great article. Hopefully it will help the masses understand why Anderson is wrong. It also leads to a worry I have had about Google. They don't really understand science at a fundamental level. Maybe I am wrong, but I get that impression all too often. - Deepak

Although I admit I don't quite get the cloud part. It almost seems incidental - Deepak

Hey Deepak. Don't tell Attila you don't think Google gets science. ;-) - Mr. Gunn

It's too late. :) - Attila Csordas

There's a good "rebuttal" on ArsTechnica (see comments to Wired piece for a link). Also discussion on Nature Network Cancerevo blog (on the computing clouds) and at the Researchers and Web 2.0 group which you are all warmly welcome to join (I sent round an invite just now with the link). - Maxine