Social Bearing should return all tweets, retweets and replies that match a given search term providing that tweets are 6-9 days old and the search term or hashtag is not a high volume search or trend.
The open search and user search will return that latest 100 tweets and 200 tweets respectively but you can load older tweets up to the past 6-9 days for an open search, or up to 3,200 tweets for a user search.
In addition, Social Bearing will not filter any returned tweets, apart from sensitive tweets. Filtering of tweets is down to your preferences, which may mean the tweet wall will be filled with popular duplicate retweets or direct replies for example. We felt this best reflected Twitter activity around any one subject and gives the user more choice.
When it comes to popular phrases or trends, Twitter search API will limit the number of returned tweets. Tweets get limited which seems when the tweet rate for the searched term approaches 100 tweets per minute. Various tests suggest that all tweets are returned when the tweet rate is closer to 60 tweets per minute or slower.
For high volume terms, with 1000’s tweets per second, it will be unlikely that a specific tweet will appear in the results. Tweets for popular trends should be treated more like a snapshot of activity rather than an exhaustive list, with popular retweets much more likely to appear.
Only access to the twitter firehose would potentially allow all tweets within high activity to be displayed. Even if we were able to display 1000’s of tweets a second, this would present many challenges in how tweets should be displayed and filtered.