Twitter annotations: the next big thing
Nearly two months ago, Twitter hosted Chirp, its first developer’s conference.
While much of the world eagerly awaited Twitter’s monetization plans, the company made an equally important announcement: annotations.
The seemingly trivial feature enables developers to add metadata to tweets created in applications, opening the door for even more Twitter mashups than we have already seen. A big deal, considering the company has been in the driver’s seat when deciding what information is attached to each tweet. Now, however, the company will be handing that role over to third party developers.
For instance, Twitterfall could add what browser the tweet is coming from. Last.fm could embed a link to the song that was being listened to at the time of a tweet. Or before sending a tweet, Tweetie could ask for your mood and attach that to your tweet.
Individually, the data doesn’t amount to much. But when collected over thousands and thousands of real-time tweets, the information can be a powerful dataset, taking the pulse of the world.
If the feature lives up to its potential, it could be a game changer for publishers, making available the tools needed to finally harness the real-time web.
How annotations work
We all know that tweets are 140 characters, but underneath the hood, tweets are actually much longer.
In most Twitter clients and on Twitter.com a tweet usually shows an avatar and the content of the tweet. For example in Tweetie, a typical tweet looks like this:
Great for humans, bad for computers. Each tweet is actually an XML file with over a dozen of modifiers and metadata. When we read our friend’s tweets, the XML file is being translated into the simple tweets we are used to. Inside, there is information such as the date the user joined Twitter, where they are located, whether you are following the user and other information. See below:
<status> <created_at>Tue Jun 01 17:06:08 +0000 2010</created_at> <id>15191066402</id> <text>Testing some Twitter API stuff</text> <source><a href="http://itunes.apple.com/app/twitter/id333903271?mt=8" rel="nofollow">Twitter for iPhone</a></source> <truncated>false</truncated> <in_reply_to_status_id></in_reply_to_status_id> <in_reply_to_user_id></in_reply_to_user_id> <favorited>false</favorited> <in_reply_to_screen_name></in_reply_to_screen_name> <user> <id>9897482</id> <name>Sean Blanda</name> <screen_name>SeanBlanda</screen_name> <location>Philadelphia</location> <description>Editor for @emediavitals / co-founder of @TechnicallyPHL. Google Reader shares addict. @BCNIPhily organizer.</description> <profile_image_url>http://a3.twimg.com/profile_images/847625399/twitter_bigger_normal.jpg</profile_image_url> <url>http://www.seanblanda.com</url> <protected>false</protected> <followers_count>796</followers_count> <profile_background_color>ffffff</profile_background_color> <profile_text_color>333333</profile_text_color> <profile_link_color>a1a1a1</profile_link_color> <profile_sidebar_fill_color>e0e0e0</profile_sidebar_fill_color> <profile_sidebar_border_color>000000</profile_sidebar_border_color> <friends_count>594</friends_count> <created_at>Fri Nov 02 20:27:45 +0000 2007</created_at> <favourites_count>2</favourites_count> <utc_offset>-18000</utc_offset> <time_zone>Eastern Time (US & Canada)</time_zone> <profile_background_image_url>http://a1.twimg.com/profile_background_images/10462374/logo.jpg</profile_background_image_url> <profile_background_tile>false</profile_background_tile> <notifications>false</notifications> <geo_enabled>true</geo_enabled> <verified>false</verified> <following>false</following> <statuses_count>3086</statuses_count> <lang>en</lang> <contributors_enabled>false</contributors_enabled> </user> <geo/> <coordinates/> <place/> <contributors/>
That data can be used by computers to combine multiple tweets to make practical use of the hoards of data streaming through Twitter every second. For example, a mashup like Trendsmap takes trending topics, combines it with the geolocation data made available in tweets and lays it over a Google Map.
Annotations will allow developers, for the first time, to add additional fields in that messy XML file you see above. You might see information like <weather>Sunny</weather> or <party>Republican</party>. Machines can then read this data and filter the tweets accordingly. So by adding the two fields above we can measure which party prefers which kind of weather. Or, as Marshall Kirkpatrick of ReadWriteWeb pointed out: imagine measuring the different in tweet sentiment between Republicans and Democrats and getting an alert every time that sentiment shifts drastically. Imagine overlaying that data over a Google News timeline to measure what events caused the most shift in our nation's opinion of the president.
What it means
140 characters will actually mean 140 characters:
Now tweets are often cluttered with retweet data, @replies and hashtags that have little to do with the actual content of the tweet. Metadata can attach that data to the tweet without sacrificing those precious 140 characters. Twitter’s retweet function uses this method to free up writing space.
As Twitter CEO Ev Williams wrote on his personal blog:
Inaccurate attribution is possible in any medium. But in Twitter, because of the character limit, it's often necessary. People shorten and edit retweeted tweets to make them fit along with the extra metadata. Even when for legit purposes, that can be misleading and unfair to the author. Worse, RTs can actually be easily faked, which has become a form of spam, wherein well-known people are shown to be promoting something they never twittered about.
Less URL shortening
It didn’t take long for Twitter users to begin using URL shorteners like bit.ly to save valuable character space when tweeting a URL. A new “URL” field in the annotations could place a small link under the tweet so users have more space to talk about the link and don’t have to bother with deceiving URL shorteners.
Mashups, mashups and mashups.
The most exciting potential for annotations is the likely explosion in new ways of mashing up Twitter data. Like the political mashup mentioned above, the possibilities are only limited by the imagination of developers.
Just as trending topic can reveal what the world is talking about, trending annotations will reveal what annotations are being used the most. For example is the “song” annotation is trending, Twitter users are listening to a lot of music. Now, users can not only see what everyone is talking about, but also what they are doing and where they are doing it.
According to Twitter, annotations are due in Q2 and there are signs that the feature will be released soon. Last week, the company held a hack day where a group of developers created new applications using annotations in under two days. The list of applications is available here. Twitter has also released a list of “recommended annotations” to get your developers started.