bit.ly and teaching the machine
It's been a while since my last post. Since then I have been learning Python and Django
. I definitely prefer Python to PHP. What I love about Python (and using Django) is that it allows a novice like myself to make relatively complex applications without frying my brain. One of my first projects with Django was to make a url shortener.
After I made a simple prototype, I began thinking about what really happens when someone shortens a url. Sharing is the only reason to shorten. It is inherent in the nature of shortening. The sender is making the "package" of the content more palatable to the recipient(s).
Think of the sending of links from person to person as a flow of information. Knowing who is sending what to whom is powerful. This information reveals shared interests between different people--the heart of social bookmarking. It also shows relationships between people, places, things, and ideas--the heart of the semantic web
. However, this flow is largely unharnessed. For example, sending links via instant messenger, e-mail, and Facebook go unnoticed by the "machine" we call the Internet. Twitter seems to be changing that. Twitter has become a hub for link-sharing.
One site, Twitturly
, uses spiders to track all of the links shared on Twitter. It then ranks the most-shared links on a Digg-style home page. The theory and execution of the site are pretty slick.
However, more importantly, by limiting "tweets" to 140 characters, Twitter has created a huge market for link shortening.
is a rising star among link shorteners. Registered users can keep track of their shortened links as well as how much traffic their shortened links have garnered. It caches each page it links to as well as processes each page using Reuters' Open Calais technology. Open Calais automatically extracts information about entities such as people and places from unstructured text.
On Oct. 31, Bit.ly posted
the day's most-visited shortened links. The links aren't too special (one is a streaming episode of the show Dexter), but the idea is powerful. Bit.ly is owned by Betaworks
, whose Summize
search engine was acquired
by and integrated with Twitter. It is not unreasonable to think that Twitter will soon make Bit.ly its default url shortener. If Bit.ly can get a vast number of shortened links, it will be sitting on the type of data that can put it above the ranks of Delicious and Digg. Just think what one can do by combining raw traffic stats with the number of people shortening (think sharing) a given url.
The problem with Delicious and Digg is that users can only send links (in a meaningful way) to other registered users. It doesn't tap into the flow of sent links. It creates its own, smaller flow instead. Delicious and Digg are unable to track url popularity beyond their own borders. Bit.ly on the other hand, taps into the already-existing flow of shared shortened links. It doesn't matter how a Bit.ly url is sent. Bit.ly operates within an infinite universe, so to speak.
Bit.ly reminds me of another interesting web service, RECAPTCHA
. In a nutshell, RECAPTCHA provides a utility to one group (spam prevention) while simultaneously providing utility to another group (using its data to refine text recognition software). Similarly, Bit.ly provides the utility of link shortening with the utility of social bookmarking with automatic semantic tagging.
Bit.ly has the potential to be one giant leap forward toward the advent of the much-hyped semantic web.
05 November 2008