Nine Years of 140 Characters
by Matt CholickTwitter recently announced they're testing a 280 character limit. There's a graph in the post that shows 9% of English Tweets have 140 characters. This was a surprisingly high percentage to me; it takes me several edits to hit the limit exactly, and I didn't think that many people went to the effort. I recall a few instances of looking up things like unicode ellipsis so that I could sneak in under the limit.
Curious how my own content stacked up against that average, I downloaded my Twitter archive and then wrote some Python to parse my tweets. Here are a few simple stats (excluding @replies and retweets):
Year | Tweets | Mean Length | Length ≥ 135 | Length = 140 |
---|---|---|---|---|
2017 | 63 | 113 | 41% | 21% |
2016 | 109 | 113 | 36% | 18% |
2015 | 108 | 112 | 34% | 12% |
2014 | 123 | 111 | 35% | 18% |
2013 | 139 | 108 | 29% | 15% |
2012 | 252 | 95 | 20% | 11% |
2011 | 281 | 78 | 10% | 6% |
2010 | 211 | 71 | 3% | 1% |
2009 | 457 | 71 | 2% | 0% |
2008 | 358 | 78 | 8% | 3% |
2008 was a long time ago, and I definitely used Twitter differently (as the table shows). Here are a few representative tweets:
I'm about to watch the best 30 minutes of animation ever created: Futurama's Roswell That Ends Well.
Interesting wired article: https://www.wired.com/2008/04/ff-wozniak/
The dude abides.
Is it obsessive compulsive to dump out a bag of skittles and sort them by color?
Java has tainted me - I like xml now... it almost doesn't seem too verbose.
They content is very random. Reading back over them, I can empathize with people advocating for the right to be forgotten. Tools like Twitter came along when I was old enough to know to not act too stupidly online, but still.... those early Tweets are so banal. They're thoughts that should be ephemeral, but instead sit there, frozen there in amber, until the end of time. Those early years had a very low signal-to-noise ratio.
Wind the clock forward nearly a decade, and I'm definitely making much better use of all those characters. Here are a few recent and representative tweets:
Deep neural network generated text-to-speech: https://deepmind.com/blog/wavenet-generative-model-raw-audio/ The audio samples near the end are amazing.
A sobering post by @briankrebs, 'The Democratization of Censorship' https://krebsonsecurity.com/2016/09/the-democratization-of-censorship/ Great to see @google step in w/ Project Shield.
Sad that I discovered @LuckyPeach only months before its demise.
Vendoring & GOPATH create enough pain & annoyance that every time I step away from Go & return, I seriously consider dropping the language.
Crazy bimodal distribution of reviews: https://fivethirtyeight.com/features/al-gores-new-movie-exposes-the-big-flaw-in-online-movie-ratings/ Wonder how often distilling movie to mean score obscures this kind of detail?
Nine years later, I think those examples are about as good as the medium can get in 140 characters. It's enough to let a folding magazine's writers know that they'll be missed. It can give enough context to a link so that a reader knows it's worth clicking. It's possible to complain about a single, concrete thing (GOPATH...) in hope of soliciting links to a solution. It's enough to broadcast a position on some contemporary issue.
I don't think Twitter is broken. I just think it's not a medium that can ever facilitate discussion: it's a mistake to think that discussion there is even possible. Nor is nuance possible. It's a broadcast medium for simple thoughts, links, and photos. 280 characters won't change that.