Recent news reported that Google has become accredited to register and sell web addresses. In the same news Google announced that they are not planning to sell domain names. What are the reasons for Google to become a web registrar then? Here is my guess: fresh URLs. When I was working at a major search engine in Germany (which is now offline), one of our main problems wasn’t the crawling or indexing of web sites - we got that part figured out and did it well - it was finding initial starting points (URLs) for our crawler.

As Fravia has taught us, a large percentage of all websites are not indexed by search engines. Commercial sites (and spammers) make sure that their web pages are known by as many web crawlers as possible. They use "submit a link"-forms, announce their sites via mailing lists and spray paint them on your neighbor’s garage door. On the other hand smaller sites (the so called "long tail" of the internet) rarely make this effort. These pages are the homes of special interest groups, non commercial organizations, communities, fan pages, obscure research projects or plain silliness - in other words: they probably contain interesting stuff. Occasionally it takes quite some time until a search engine discovers these sites, if they are discovered at all.

Being a registrar Google should get access to the master data, which contains a list of all registered domain names. This information could be used by Google to broaden their sources. Note though: this is only my take, I might be mistaken.