Fjord Philosophy

When a customer of a hosted software company like Dropbox or Google Drive stops paying their bill, it’s reasonable to expect that the service will delete their stored data after a certain amount of time. It costs the service money to keep it around, and we couldn’t expect them to steward it without compensation indefinitely.

I think of this situation as being similar to a defaulted account at a self-storage facility. When the grace period has elapsed, the terms of the facility’s agreement allow them to stop storing the unit’s contents. In some cases, that means they’ll sell the unit off. What was once private becomes public during an auction, or it becomes private to someone else if the facility sells it directly.

Thankfully, doing the same thing in the realm of digital storage would be considered wildly inappropriate and, in many services’ terms of use, illegal. (All other sorts of data use and abuse notwithstanding.) It would be shocking if, for example, Flickr put your photo library up for auction if you defaulted on your Pro membership.

But consider the somewhat unlikely scenario where a service is storing culturally important information, like only-copies of works of art or documents of historical significance. It might seem far-fetched, but it happens. Think of the countless musicians who store unreleased music in Dropbox, or the journalists whose manuscripts, with all the original Track Changes marks, are in Google Drive. Even corporate tenants’ files could be considered culturally significant if they pertain to an historic event or an innovation. Would we want those to go away just because somebody stopped paying their bills?

In the case of a physical storage locker, it’s easier for facility managers to intervene before valuable things are destroyed. They can see the priceless folk instrument or the classic movie prop’s silhouette beneath a dust cover. Maybe more importantly, the facility manages a relatively small amount of units, and the rate of defaulting is slow, so they’re more likely to have the time to review units and decide what to do with them.

But that’s not the case for platforms that store digital objects. They might store so much stuff that it would be infeasible to individually inspect a customer’s data (taking for granted, for a second, that we’d want them to!). And even if it were, it might be virtually impossible for the managers to distinguish that Important Stuff from all the other data that ends up getting stored — no parchment paper or collectible baseball peeking out from the corner.

A big caveat here is that digital artifacts are not perfectly analogous to physical artifacts in that: they can be flawlessly duplicated an unlimited number of times; they don’t bear the individual physical marks of their creators’ manipulation, as in a painting’s brushstrokes; and they’re often — let’s be honest — valueless. Even a storage unit’s Beanie Baby might be considered more worthwhile to save, given our shared values, than a digital storage tenant’s database of 14 thousand archived invoices. (On top of all that, let’s not forget that the physical storage facility’s model of “saving” important things — selling them — hardly guarantees that the contents will be stewarded well.)

We might also say: there’s no problem here, because people ought to know not to store really important things in cloud services. They should have local backups on multiple media, etcetera. But people do store important things in cloud services, without redundancy. With PSAs about digital hygiene, we could make incidents less common, but we probably couldn’t eliminate them.

This seems like a problem to me, because we want to be able to step in and save important stuff from destruction, but it’s an obvious and seemingly intractable violation of users’ privacy to snoop around in their stored data just because they haven’t paid their bill, or for any reason. And that’s taking it for granted the the service’s architecture even allows for such snooping the first place; many data models use encryption or other barriers that make snooping impossible.

So: what do we do? Do we tell people not to store their Great Works or historical records in cloud services? Do we make public structures for storing data in trust? Or do we accept that a certain amount of loss is inevitable? It might seem crazy to care about these things falling through the cracks when the net we’re currently casting for information preservation is so wide — on our current path, future anthropologists and historians and who knows whoever else will not be wanting for records of this time. But it is sad to think that a self-storage operator in Bloomington, Illinois could save a rare Eames chair from destruction, but Dropbox couldn’t discover and steward the unreleased photographs of the next Vivian Maier.