After all the "misinformation flying around" social media that ended up as a part of the official news record during last week's quick moving hunt for the Boston Marathon bombing suspects, it would make a lot of sense for Twitter to implement a button that lets you edit tweets. Even President Obama said as much: "In this age of instant reporting and tweets and blogs," he said Friday after the capture of Dzohkhar Tsarnaev, "there's a temptation to latch on to any bit of information, sometimes to jump to conclusions." So wouldn't it make sense to be able to jump back, correct the record, and get false information out of Internet circulation? Well, yes, but it's unlikely Twitter would make such a move, according to two former Twitter developers and three social media developers who spoke with The Atlantic Wire Tuesday.
The two most damning falsehoods of the Boston frenzy—the unapologetic New York Post cover and the misbegotten outing of a 22-year-old missing Brown student—left many real-time enablers regretful, including many powerful Twitter users. Perhaps the most proactive among them is Wired's Mat Honan, whose admittedly offending tweet is pictured above at right and who is openly advocating for an Edit button today, even offering a specific solution:
So here’s one way it could work using Twitter’s metadata. Twitter could add a function, similar to a retweet or favorite, that let you edit and correct a tweet after it had been posted. Those tweets then show up in a timeline as having been corrected–again, they could be flagged like favorites or retweets. Click on a tweet marked as edited, and it uses Twitter’s Cards function (the same system that lets tweets embed images, videos, and text) to show the original.
Honan also suggests that Twitter give the original author the option of notifying everyone who retweeted the offending tweeter's misinformation of the correction, which they could then choose to republish at the top of their timelines, if they so chose. He's not the first to call on Twitter to build in corrective functionality, but developers who actually know how to build Twitter say it's both technically and practically impossible. "No, I can't imagine that they'd implement that in the foreseeable future," Blaine Cook, a former lead developer at Twitter, told us this afternoon. Twitter doesn't have "anything to share on this" mountain of dev work, a company spokesperson told The Wire, but it would be an uphill—and perhaps unnecessarily complex—climb. Here's how:
The Technical Problems
Just think of all the places a tweet will go: Your feed, your followers' feeds, then—if it's retweeted—to their followers feeds. And then there are all the apps and clients that serve all these functions on top of that. Going through each feed and inbox to change that data attached to each message would be a "dicey prospect," says Evan Prodromou, founder of Status.Net, an opensourced social service. "Keeping modified tweets in sync—across all the clients running in the world—would be incredibly difficult and expensive compared to just pushing out a live stream, which is already difficult and expensive," said Cook.
To add this functionality, Twitter would essentially have to upend itself. "Introducing editing would make Twitter's architecture far more complicated, particularly when you consider that a user posting a tweet results in a cascade of activity across different internal systems," Alex Payne, a former Twitter engineer who worked at the company from 2007 to 2010 told The Atlantic Wire. Twitter works as quickly as it does because the platform runs on multiple parallel servers, a developer here at The Atlantic, Chris Barna, explained: "The system that handles the request creates a list of 'tasks' that are independent of each other that can then be handled by many servers," he said. These independent requests, however, assume that a tweet won't change depending on the server or the order of the request. "Modifications to tweets would increase the difficulty of maintaining search, advertising, and other features of the service," added Payne.
In addition, to update any given stream, Twitter relies on a lot of caching. Adding an Edit button—unlike, say, the Delete button—would add multiple layers of complexity to that data load. The system would have to look and see if any tweets have been updated... and then go back again to look and see if any retweets had been updated. Dalton Caldwell, the CEO of App.Net, explained it in terms of editing a YouTube video: "When I upload a video to YouTube it goes out and its permanently there. If I edit a video, what actually happens is a brand new upload, and YouTube just updates its pointers," he told us. That's not only a tricky maneuver, but it could complicate things to the point of not working. "Caching relies on systems to be as simple as possible so that caches aren't invalidated too frequently," Barna said.
The Practical Problem
Beyond the technical headaches, adding an Edit button would change the way we interact with Twitter. It's a lot of woulda-coulda, really. Right now, tweets are static; we treat them as such. "The widespread conception of Twitter is based on ephemeral, one-time communications," said Payne, the former Twitter engineer. We assume something we retweeted will never change. With an update or editing function, we would be subscribing to all of Twitter's maleable new realities. What if we don't want to do that? That question would linger in our minds every time we RT something. Plus, it would be a pain to see if your RT was still legit. The old "RTs don't equal endorsements" adage, well, that would be upended, too.
From an interface standpoint, Twitter would have two options: It could delete any edited tweet from our timelines, but then a user might wonder where that tweet went—or might not even notice. Alternatively, Twitter could give users a way to retweet a corrected tweet, if they wanted to. (Honan suggests the second option.) Building out that interface is a project in and of itself, but it's also a different version of Twitter than the quick moving network that we know today.
Initially, a new Twitter Edit button would probably take a lot of explaining to Twitter's 200 million-plus users. "How would you explain what an addendum tweet is?" wondered Caldwell, who thought Honan's solution, although complicated, makes the most sense. People could get used to editable Twitter, if designed well. But truly modified tweets would change the ethos of a social network from what's happening right now to, well, to some sort of constantly updating blog post you can check in on, and that's kind of not the point.
Conclusion: A Twitter Edit Button's Not Really Worth It — to Twitter, or to You
Considering all the work and resources it would take for Twitter to build a Twitter that you can edit, there's not a lot of payoff. Seriously: How often does this kind of mass mistake happen where it really, truly matters? How many average users affect average people with a tweet they can't just delete? In the hunt for the Boston bombing suspects, an average family already grieving over their missing son had his name connected with a deadly bombing—by above average Twitter users, to the rest of the world. That's not the norm. "Does building the UI and the infrastructure to allow editing tweets pay off for Twitter? I think probably not," said Status.Net's Prodromou. Indeed, to Twitter, it would probably save a lot of time, money, and energy if we all just tweeted with a lot more care.