I’m confused by this article about preserving digital science publications:

Avoiding a digital dark age

“If all works out nicely, this represents a golden age for researchers: unlimited new online collections of data and research information with powerful tools for aggregating, analysing, and accessing that information. But what are the risks?

Being able to preserve digital data is a must for a golden age of research information, and a major risk is therefore the rapid obsolescence of digital objects….[Without stewardship and continuity of access] these online research collections and datasets will never last long enough to revolutionise the way we do research. At worst a new digital dark age will follow where access to the previous generations’ information is severely compromised….”

According to Open Access News , the author is “the Project Leader for the Australian Partnership for Sustainable Repositories”, so I guess he should know what he’s talking about. But facilitating a scientific revolution seems to be beside the point. Are scientific revolutions really built on access to old publications? I would have thought that most really active scientific disciplines can and do mostly reference the past decade or two’s research. The joke goes that digital information (as opposed to that which is printed on paper) lasts for eternity or 5 years, whichever comes first. But that’s more true of stuff sitting on single disk drives. Institutional repositories are probably good for at least a couple of decades. Longer, assuming society stays stable and there is a reliable lineage of people working to get the repositories upgraded to the best available open-standard file formats and imported into the ever-expanding storage pool. That kind of basic action seems likely to happen under the status quo.

What has always worried me is what happens to science, published in the digital age, if society goes unstable. We have access to pre-enlightenment research mostly because books are durable, probably more durable than rack-mount storage servers and the software that runs them. But if we did have a little hiccup, it seems possible every journal not hard-bound and stacked (and the rise of open-access journals seems to be synonymous with the rise academically significant electronic-only journals) would go poof.

Strikes me that that’s the real concern for digital preservation. Perhaps I’m being millenarian (heaven forbid). Still, I think the APSR should get in touch with these guys. Micro-etched nickel spheres. That’s where it’s at for journals who want their knowledge to span the inevitable society-collapse-gap.