This is documented, supported and subject to as much change or stasis as any other Twitter feature.
The entire tweet is given to avoid an extra round-trip in rendering timelines. Many our results are denormalized in this way, as a fully normalized schema delivered via an Internet service would be impractical from a display latency standpoint. -John Kalucki http://twitter.com/jkalucki Services, Twitter Inc. On Wed, Dec 2, 2009 at 7:31 PM, M. Edward (Ed) Borasky <zzn...@gmail.com>wrote: > I've been looking at data acquired from the "sample" feed. In those > tweets, I've discovered something, and I wanted to run it by Twitter > and see if this is an official, documented API feature. > > When a tweet is a retweet using the retweet button on the web page, > there is an extra key-value pair in a returned tweet. I've been > handling them as JSON objects (in Perl), but I suspect other formats > do the same. The key is "retweeted_status", and the value is an > *entire* object containing the original tweet that was retweeted. If a > tweet was not a retweet, this extra key-value pair isn't there. > > Questions: > > 1. Is this official, documented behavior, or did I discover it on my > own? > 2. Is it likely to change? > 3. Why is the *entire* original tweet given as the value? Wouldn't it > be enough to give the tweet ID and let the app go hunt for it if > necessary? >