Jason Scott’s blog post from a few weeks ago, Eviction, or the Coming Datapocalypse, has kicked up a bit of dust. His argument (see also his follow-up post) is that services for hosting user-generated content need to take more seriously the consequences of shutting off those services. He argues that shutting down a hosted service should carry similar requirements as evicting a tenant: notice, minimum delay between notice and lockout, etc. In his follow-up, he proposes that regardless of the law, rogue archivists ought to look for closing-down services and spider them before they go down.
A couple comments, in no particular order:
- Scott makes a very forward-looking argument. There’s value in digital content — not just personal value to whoever made it or visited it in its earlier incarnation, but social value. Social scientists of all stripes may be interested in the content as artifacts of human communication. That’s why we have archives of any variety.
- The eviction analogy applies equally for particular users as for the wholesale shutdown of a service. If your account will be shut down for whatever reason (e.g. because you are accused of violating the terms of service), your data should be saved for a minimum time for you to retrieve it. (For public-facing services, it’d be acceptable to remove the content from public view.)
- If you were inclined to legislate in this area (see below), there’d be an easy hook to do it: Section 230 of the Communications Decency Act, or Online Copyright Infringement Liability Limitation Act of the DMCA. So you say you host user-generated content and should have a safe harbor from responsibility for the content? Okay — but in order to qualify for limitations on liability, you have to follow some guidelines to protect users and preserve the historical record.
- You could use the same hook to legislate on other issues of consumer protection for UGC sites, such as privacy or data portability.
- So is there an argument for legislating here? I think there might be. “The market” may underserve users (and historians) here, because by the time a company has decided to shut down a service, it’s pretty much turned its back on those customers. It’s the same issue with eviction: once your landlord has decided to kick you out, he’s probably not planning to do more business with you. In both cases, there’s little incentive to treat the consumer fairly.
But on the whole, this kind of behavior will erode consumer confidence in these services. So will companies offer (and abide by) more favorable terms, and will consumers choose services that do? Maybe — but there are collective action problems, bounded rationality, and monopolistic competition (network effects, etc.) to contend with. So the market may not be good at solving this problem for users, to say nothing of the public good of historical preservation.
- What exactly should the service provider’s responsibilities be? At a minimum, the procedures should provide users with a reasonable opportunity to retrieve their data. For public-facing services (e.g. Web hosting, public photo sharing, etc.), there should also be a reasonable opportunity for archivists to make preservation copies. In both cases, there are significant questions of utility: in the former, the question of portability (OK, you have your data — now what? Is it in an open format? Can you take it elsewhere?); in the latter, the question of access (all this stuff is presumptively copyrighted — how can it be made available?). There’s also the more immediate question of how to enforce such a policy (whether it originates in private contract or public law).