My friend, neighbor, and colleague Bob Frankston has been railing about the limitations of the Internet's Domain Names System (DNS) for some years, and I am starting to see things from his point of view. Bob's argument is that while DNS provides the valuable service of mapping domain names (and hence URLs) to the lower-level IP addresses, that domain names themselves are too limiting. They might have worked well when there were a small number of easily remembered and spelled domains, such as microsoft.com and apple.com, but it is increasingly difficult to find a short, mneumonic name that isn't already taken. I thought Bob had a point, but I also appreciated the convenience of seeing a URL on TV or on a billboard and being able to recall it hours later when I was in front of a Web browser. Bob's proposal that web browsers incoporate a directory service on top of DNS appeared to add an extra level of complexity. Besides, with a name like Herot, getting the domain Herot.com wasn't all that difficult
More recently, I'm starting to see Bob's point. My epiphany came while choosing a name for a new business we are starting at Convoq. Picking the name was easy compared to registering the domain. It seems that every word in the dictionary is registered as a .com, along with a surprising number of two-word combinations. Invariably, the owners wanted somewhere in the deep six figures to part with it. For the most part, the actual sites were just a collection of links that bore some vague resemblance to the name of the domain - the kind you may have stumbled upon when you mistyped a domain you were looking for.
Some recent news stories have shed some light on this phenomenon. In Millions of Addresses and Thousands of Sites, All Leading to One the New York Times describes how NameMedia has raised $120M in funding and owns 725,000 web sites, all populated with generic content. It seems that 5 to 10 percent of people searching for something in the web will type a word into the address bar instead of using a search engine. Even though a lot of these words only generate four or so hits a month, it adds up in volume. Some of those sites, like photography.com have custom-generated content. Others just show the same links that one would get if one typed the same word into a search engine, and generate revenue for that referal if someone clicks through.
According to a recent article in Business 2.0 trading in domain names has become a brisk business, with names such as greetings.com going for $350,000 at a packed auction in Las Vegas. The article describes how Kevin Ham set up a deal with the government of Cameroon to share the ad revenue when people type .cm instead of .com and end up on one of Ham's sites via a wildcard match he set up.
The ultimate irony is that when every word in the dictionary leads to a domain, the address bar becomes a search engine itself, and with most of the words owned by search engine companies, users will need a search engine to find the actual content they are looking for. Bob's vision for directory services replacing domain names will come about, although not quite the way he imagined it.