(Sort of like xkcd)
Something is happening to science today (not the web site "Science Today"). Postmodernism is undermining the meaning of truth - some scientists are now saying truth is whatever the consensus is. Similarly, the principle of "reproducibility" is under attack.
An interesting article at Ars Technica.
Their concern is largely on the software tools (encouraging the adoption of open source), which is a good point.
There is a second concern (which I heard expressed on comp.arch). If you need a one-of-a-kind machine to generate results (a supercomputer, particle accelerator, whatever), then you have immediately sacrificed reproducibility.