Microblogging site Twitter is serious with its fights against spam. To tackle this problem, Twitter developed BotMaker to be its very own anti-spam fighting system.
The BotMaker is an anti-spam system that is designed to reduce spam that a user sees on Twitter. Developers of the BotMaker followed three principles as their guide to design the program:
(1) Prevent spam content from being created.
(2) Reduce the amount of time spam is visible on Twitter.
(3) Reduce the reaction time to new spam attacks.
Preventing creation of spam earlier will result to the reduced spam a user sees on Twitter. If there's spam, Twitter will aim to reduce the amount of time spam is available on the site. And lastly, responses to spam must be reduced by decreasing reaction time. This also means Twitter acting on an issue or spam report as quickly as possible.
Twitter inspects data and events coming from its distributed systems, following a set of rules. Once the problem is identified, Twitter will act according to the guidelines.
Twitter bots have two parts: one part that decides whether to act or not on an event and the other one, actions that tell what the caller must do.
Twitter faces a big challenge in supporting this system because the rules can run on Twitter's other features like Tweets, Retweets, Favorites, Messages, and Follows.
Fighting spam messages is a challenge for Twitter but the company is determined to reduce spam. Twitter said that since BotMaker rolled out, spam was reduced by 40 percent.
Raghav Jeyaraman, Twitter engineer explained in blog post how Botmaker will stop spam.
"BotMaker language is type safe, all data structures are immutable, all functions are pure except for a few well marked functions for storing data atomically, and our runtime supports common functional programming idioms," Jeyaraman wrote.
The BotMaker simply breaks down spam messages into real-time, near real time, and batch jobs for easier categorization. Twitter introduces Scarecrow, a tool that spots problem URLs and account name even before a spam tweet is written and posted. The Botmaker also has the Sniper tool that searches for spam messages Scarecrow may have missed. And lastly, batch jobs that analyze large amounts of online data to identify long-term behavior patterns. Doing this can help in making smart online models.
Twitter will constantly work to fight spam. While there is a possibility that spammers will eventually bypass Botmaker, the company will still try to find more efficient ways to identify spammy events even before they are shared on Twitter.
