The social nature of the shortened URI
Right this minute, there's a curious problem with the Blogdex front page: Many of the links go through TinyURL.com. This isn't really a bug, since those links do in fact go where they're supposed to. But when Blogdex shows those obfuscated, shortened URIs instead of where the content actually lives, the services becomes notably harder for a reader to sift through.
To be fair, this isn't a common occurrence. Blogdex appears to have been accidentally spammed by airbeagle.com, which has a lot of subdomains that Blogdex is counting as separate blogs even though they have identical content. Still, the appearance of inscrutable addresses adds a further wrinkle to the question of when a domain name might matter.
Last year, I wrote that services like TinyURL.com illustrate how, as the act of manipulating links gets increasingly more convenient, the dissemination of those links takes place along increasingly personal lines of transmission. If your friend sends you an email with a strange URI, for example, you'll probably trust that it goes where it's supposed to because of the friendship. (Unless your friendship is one that includes URI pranks, at which point you're still trusting that your friend won't send you any link that grosses you out too much or gets you fired.)
But when you don't have any personal relationship with whoever recommended the link, you're less likely to follow that link. For years, Slashdot posters had to be careful about clicking on links in user comments for fear of being sent to http://goatse.cx. In most cases, though, the risk is more prosaic: That of having your time wasted by a page you don't care about. As a Blogdex reader, I depend on the domain name in a URI to tell me a little more about what's being pointed to. Is this an article on the Washington Post? Or another annoying non-story on the Drudge Report?
(As a side note, sites like Blogdex and DayPop would be even more helpful if they included more information about each link. Having to mouseover links in Blogdex to see what text people use to link to the result is slow, and it would be, like, totally awesome if the result could also include some sort of excerpt.)
There are a lot of different ways to pass around URIs, and they range from the personal to the impersonal. Maybe you email or IM a link to a good frend, or you email a link to a mailing list you're on, or you post it on a blog or a wiki, or maybe it's even picked up by Blogdex.
As creating and distributing hypertext becomes cheaper and quicker, and communication continues to decentralize, more and more recommendations of URIs will depend on implicit factors of trust, reputation, and personal histories. I read a lot more online than I did 5 years ago, and most of that reading is guided through sources I have some sort of relationship with: Blogs I regularly read, friends who know my interests, etc. Perhaps it would be worthwhile for services such as Blogdex and Google to try to infer the influences of these implicit social factors. It may not be enough to ask: Who is pointing to this page? It may be even more useful to ask: Why are they pointing to this page? Who do they think would enjoy reading it? And are you one of those potential readers?