DarkMagus wrote:
In a statement on its website, the CRU said: “We do not hold the original raw data but only the value-added (quality controlled and homogenised) data.”
If they aren't being forced to recreate their homogenised set (as in there hasn't been anything put forward by the scientific community that completely blows away the initial axioms they drew up when they created the aforementioned set) then this is not an issue. Anyone with a background in IT, and/or science, should know this. You only need to go back to your raw data when your whole analysis approach has to change. Or you were remarkably dumb and got the granularity of a series of variables that you absolutely need at the wrong level (which can happen of course, especially in a research lab speaking from experience). Also, as Merrick points out, the raw data in question does exist elsewhere too.
I did reply to your other post too btw.
http://headheritage.co.uk/headtohead/u_know/topic/55494/threaded/695396
Edit : And yeah, there is a lot of shit code running in research labs. It does annoy and horrify me somewhat. Scientists have a habit of thinking they can write code, or they think they know who is a 'good guy' when it comes to interviewing developers. It's fucked up. But then, if you see the salaries of developers in such enviroments then it isn't really surprising.
|