It’s conventional search engine optimization wisdom that URLs should contain words, separated by either dashes or underscores. This approach improves the readability of the URL – making it more usable for people while simultaneously giving internet robots something to work on.
But with people sharing URLs within places – like Twitter and Facebook (and … and … and …) – places with a default social context, we’re seeing a URL’s context trump its readability as a significant usability factor.
Who is sharing and how they describe what they’re sharing is more important than the readability of the shared URL itself.
Leaving the search engine robots blocked out completely (disallow, nofollow, etc) or piecing together a pile of redirect URLs (which may or may not exist tomorrow, e.g. RE07.US).
Additionally, the share-er’s pays for each URL with their social capital. ‘Good’ URLs (as deemed by each individual follower) raise the share-er’s capital while ‘bad’ URLs lowers.
Throw in the proliferation of other difficult to index assets like images and video – and we’re talking about an internet that’s not Search Engine Optimized, but Social Engagement Optimized.
It will be interesting to see how search engines adjust to new forms of linking. I think the reputation effects of good v. bad links has been with us for a while now through blogging so it’s not an entirely new problem. Figuring out what’s worth measuring and how to weigh that data in order to provide relevant results to searchers remains an ongoing challenge.