From the article, regarding scientific literature: "Because you're trying to present what happened in the lab one day as some fundamental truth. And the reality is much more ambiguous. It's much more vague. But this is an artifact of the pre-network world. There was no other way to communicate this kind of knowledge other than to compress it."

A statement like this suggests that the speaker either unfamiliar with the way scientific data is actually turned into papers, or inappropriately optimistic about the utility of making the data "available." It is true that scientific data can be voluminous, but the overwhelming majority of papers do not "compress" data. To stretch an inadequate analogy, scientific literature is much more akin to metadata. Imagine scientific data as a large set of digitized recordings of music, all jumbled about. The paper would represent the list of song title, artist, etc. that someone had to put together. The metadata is not so much a compression as a re-representation and categorization of the data.

As a neuroscientist responsible for sharing my results with the world, I've taken reasonable steps to ensure that all of the data used in my papers is freely available (under the Science Commons license, which I'm quite grateful to Wilbanks & co. for). Similarly, the code I wrote to extract meaningful parameters from the data and present them in an aesthetically pleasing way is also freely available. I maintain no illusions as to the utility of the database: nobody is really interested in recreating the figures in the paper from the original data, nor in reanalyzing the data. However, I do know that some of the insights I've presented have influenced those (few) that have read my papers and struggled to understand the ideas presented within.

There is nothing wrong with the idea that scientific data and biological materials ought to be readily available to those who would use them. But the notion that somehow the hard-won insights that come to those who spend years collecting and thinking about the data will somehow follow is fanciful at best. Peer-reviewed, editor-selected papers are not compressed versions that are easier to transmit, but rather the collected insights and interpretations that allow us confidence in the work we've done. So by all means, if Mr. Wilbanks can find people to pay for it, make it easy to disseminate data. Just don't be surprised to find that "decompressing" papers doesn't do all that much to advance knowledge.