British Culture Secretary Andy Burnham wants to apply movie-style ratings to websites. His plan is to have major ISPs provide child-friendly services based on these ratings. He's getting close to a good idea, but he's not quite there.
He also begins with a false premise:
“It worries me - like anybody with children,” he says. “Leaving your child for two hours completely unregulated on the internet is not something you can do."Leaving your child for two hours completely unregulated on the internet is not something you should do. You shouldn't leave your child for two hours to watch television unsupervised, either.
Here's the problem. You're going to try to separate out websites that are child-friendly, or age-range friendly, from other sites. Trying to do this via government regulation is doomed to failure. What if the content of a site changes? What about sites where comments are allowed? Take a site like Bad Astronomy, run by Dr. Phil Plait. Phil generally keeps it PG, if not G, but very occassionally he'll slip in some saltier language. The commenters do, too, but are generally good. What rating do you give this site? Who is going to validate the ratings that sites are given?
Another problem is that you're pushing the enforcement (at least partially) onto the ISPs. If they have to provide different levels of service, that involves either redundant networking or filters that run on all traffic, based on the subscriber. The redundant networking might be the cheaper option, since it doesn't require a lookup over a possibly-sparse set of subscribers for every packet or session. Plus, at the ISP level, you're probably going to be stuck with a pared-down internet or no filtering. That means keeping your children "safe" locks you out of quite a number of useful sites, even if the kids aren't even in the house.
There's a market solution here, and it could be implemented fairly easily without any government regulation or intervention. I call it .ratedg, or for the more network-geeky of you out there, RATEDG-DOM. The idea is that a new top-level domain (TLD) is created, where the sites are vetted initially and periodically for standards of content. On end-user computers, optional filtering software would be installed that only allows connections to sites in the .ratedg TLD. Traffic to other TLDs, or to numeric IP addresses, would be blocked. This isn't completely trivial, since all connections have to be addressed numerically, but a local DNS cache could handle this.
The TLD would be maintained by a nonprofit organization, which is funded by the fees paid by sites wishing to be listed. Some accommodation would need to be made for sites of nonprofits, hobbyists, and other cash-short groups. The fees could, for instance, be determined site-by-site based on the website's circumstances. This adds complexity, of course, but is not insurmountable.
The filtering could be provided by free software (funded by governments, the nonprofits, or volunteers), and disabled by password so that adults can have less-fettered access. One drawback of this is that software-based filters are notoriously susceptible to circumvention by clever youngsters. Of course, you're not letting your kids surf unsupervised, are you? The filter provides you with a no-accidental-bad-stuff net. You see the "Rated G" certification immediately, and if you click on a .com, .edu, .org, or whatever link by mistake, the filter blocks it for you with a friendly warning.
Another option, perhaps better for parents who insist on letting their children use the internet unsupervised, is a hardware filter. This would be inserted into your computer either between the network interface card and the wire, or between the card and the rest of the system (the latter is probably better). Filtering is now physically enforced by a key, which the parents can carry along with them. The case would have to be locked, as well, to keep the filter card from being removed.
This system also has extensibility. A RATEDPG-DOM could be created as well, or TLDs corresponding to other countries' standards. A hierarchy of allowed TLDs would arise naturally: a filter set for PG websites would allow both .ratedpg and .ratedg.
Note also that this would not force sites out of other TLDs that are otherwise descriptive. We already have websites that are in both, say .com and .co.uk. Having disney.ratedg would not preclude the same site being available via disney.com, though Disney would probably maintain pieces of its website other-than the G-rated portion.
Enforcement places some burden, but again the registration fees help fund this. It isn't feasible to completely verify a site's content, certainly not with any great frequency, but a combination of random compliance checks and public problem reporting can have great impact. If, say, Goggle contributes some of its resources to compliance monitoring, a much more complete picture could be obtained.
There are potential problems with this solution, as there are with any solutions. Filtering based on TLD punishes sites that do not or cannot register, and students who need access to information on non-registered sites. The Smithsonian Institution comes to mind, since there's likely to be photos of art that some would consider inappropriate for children. As the popularity of .ratedg grows, non-registered but compliant sites would feel pressure to register. Other sites might re-structure their content into G-rated and non-G-rated components, on separate IP addresses (though not necessarily different hosts). Still, other volunteer-run educational but not specifically child-oriented would likely be excluded. I'm sure there are other potential problems as well.
So that's my suggestion. It's an opt-in filtering system, both on the user and content provider ends. ISPs don't need to add or modify infrastructure. Governments don't need to add regulation or oversight. At-home implementation could be done by either hardware or software. The system is extensible. Implementation can be phased in, and some of the technical design could precede full implementation (such as prototype software that filters on .org or .edu). Since it doesn't require a strong-arm approach or extensive new infrastructure, the whole system could probably be implemented and deployed in under a year.