In a thoughtful post, Ev explains the thoughts behind why twitter implemented Retweets. There are a lot of good reasons for adapting this user generated convention into mainstream use. MG Siegler, Sean Bonner and others provide some insight into how this adaptation may have gone to far, but I wanted to explore why this feature was implemented the way it was.
Two weeks ago Twitter partnered with Bing and Google to provide the real time firehose for inclusion in their search indexes. Most likely these deals are providing the majority of Twitter's revenue to date. Form follows funding, which isn't necessarily a bad thing.
Real time search is hard, especially for sorting out noise and spam. By making RT an explicit gesture, and keeping that gesture clean by disabling it RT editing, the new RT provides a strong signal for the search index. A message that was Retweeted 15 times might be more relevant than the tweet that fell in the woods by itself. Now this is just one factor for a search index, but I'll bet it came up in partnership talks.
It will also be good for the mainstream user experience and growing the user base of Twitter itself. Most twitter users aren't going to actually tweet, but read, and that's okay. What they read won't be full of arcane syntax and there will be noise.
But I wonder if there is a way to parse RTs and mentions for benefit of search and readability, while letting users continue to edit and annotate, with a consistent user experience across Twitter and its clients.