Our minds on media.

Steven Johnson has jumped on the Web 2.0 cluetrain along with a lot of other people. He uses a great metaphor to discuss web 2.0, that of a rain forest versus a desert. I think his metaphor is appropriate for more reasons than he realizes, though, namely that this jungle, built on vast quantities of data, has no inherent mechanism for trust. In other words, in just the same way that Web 2.0 makes room for a lot of niche activities, it also makes more room for predators.

I think Web 2.0 is a promising concept. But then so was email and look at what happened with spam. Already two of the more interesting features of Web 2.0, comments and trackbacks, have been declared dead by many in the blogging community. Cause of death? Spam — that most adaptive and brutal hunter of efficient information systems. As we use the technologies of Web 2.0 to open up new streams and venues of information, this time we should take some precautionary measures against its abuse.

Comments on blogs could still be saved if more people required TypeKey style log-ins. In fact, I’ve always wondered why log-ins themselves couldn’t be part of a major API on the web. Why isn’t there a central network of log-ins a la domain name registration? You choose a registrar, register your email address with an associated password and web sites access this log-in registry through a web service API. You’d have one place to go to change your email or password. More importantly though, you’d have a distributed “white list” that would be inaccessible to bots and make it relatively easy to root out or ban abusers.

Every community needs a structure of trust. Secret societies have their passwords. Clubs have memberships rosters. Most of these examples seem antithetical when talking about the web because people think it should be inherently and totally open, but I think, when talking about Web 2.0, it makes sense to start thinking in these terms. And openness and trust are not mutually exclusive either. The folks over at Wikipedia know this and have gone to great lengths to fight garbage getting into their wiki while still allowing anyone and everyone to contribute. An information system’s worth is only as good as the worst information in it.

Another solution to the problem (and essentially a potential feature of Web 2.0 sites) is to build friendster-style trust networks into these programs and protocols. As a user I wouldn’t just search del.icio.us for any of its results. Instead I would search del.icio.us for only those links that have been tagged by people I know — or even people that know someone I know. (For that matter, your degree of separation just becomes a parameter of the search.)

Consider how Steven Johnson describes information traveling from blog to RSS to group blogs to other aggregate systems and APIs. Information that was bad or misleading on some static web page on the old web is now poisonous on Web 2.0. That just doesn’t need to be the case.