Tweeting in mysterious code
March 21, 2010 3:46 PM Subscribe
Why do people tweet in code?
So I was following tweets about the health care vote over on Huffington Post and noticed that some tweets are completely in code. For example, one tweet read:
TkJDIE5ld3M6IG11bHRpcGxlIHNvdXJjZXMgc2F5IFN0dXBhayBpcyBhIHllczsgcHJleiBleGVj IG9yZGVyIGxhbmd1YWdlIGFjY2VwdGFibGUuIEhlIHdpbGw
Why would anyone tweet in code? Is it a secret message to their cult? Are they tweeting from hardware (Mac?) incompatible with the rest of the universe? Is it a virus which is now devouring my hard drive? Or just a momentary failure of the di-lithium crystals at the heart of the Internet?
So I was following tweets about the health care vote over on Huffington Post and noticed that some tweets are completely in code. For example, one tweet read:
TkJDIE5ld3M6IG11bHRpcGxlIHNvdXJjZXMgc2F5IFN0dXBhayBpcyBhIHllczsgcHJleiBleGVj IG9yZGVyIGxhbmd1YWdlIGFjY2VwdGFibGUuIEhlIHdpbGw
Why would anyone tweet in code? Is it a secret message to their cult? Are they tweeting from hardware (Mac?) incompatible with the rest of the universe? Is it a virus which is now devouring my hard drive? Or just a momentary failure of the di-lithium crystals at the heart of the Internet?
It's encoded in base64. Try using this decoder.
Apparently, this twitter client encodes text like that to allow longer messages.
posted by Memo at 3:51 PM on March 21, 2010
Apparently, this twitter client encodes text like that to allow longer messages.
posted by Memo at 3:51 PM on March 21, 2010
Is it possible that base64 is a way to cram more information into the twitter character limit?
It doesn't seem like it based on the numbers of characters in the two messages. And the SMS character space looks to be 144 characters, so reducing that to base 64 seems silly.
Maybe it is some kind of plugin designed to obfuscate one's tweets so they aren't searchable?
posted by gjc at 4:00 PM on March 21, 2010
It doesn't seem like it based on the numbers of characters in the two messages. And the SMS character space looks to be 144 characters, so reducing that to base 64 seems silly.
Maybe it is some kind of plugin designed to obfuscate one's tweets so they aren't searchable?
posted by gjc at 4:00 PM on March 21, 2010
@gjc: no, base 64 increases the length of messages, because it reduces the character set. The original purpose was to make it safe to send in emails, etc., where filetypes and character sets were poorly implemented.
This is clearly a soft-encryption, like rot13.
posted by tmcw at 4:03 PM on March 21, 2010
This is clearly a soft-encryption, like rot13.
posted by tmcw at 4:03 PM on March 21, 2010
I wrote that before Memo's message. And by character space I meant character set.
posted by gjc at 4:10 PM on March 21, 2010
posted by gjc at 4:10 PM on March 21, 2010
Best answer: If you are talking about Chuck Todd's twitter he said PING was having a problem.
posted by govtdrone at 4:12 PM on March 21, 2010
posted by govtdrone at 4:12 PM on March 21, 2010
Response by poster: So the answer is, it's base64 encoding, and the encoded tweets are either intentionally obfuscated, or are from a twitter client that sends base64 (for whatever reason), or are due to a problem at ping.fm.
The article on base64 over at Wikipedia has an example showing how every three ASCII characters are encoded as four base64 characters, an inflation of 33% in the message size, so using base64 actually means you can send only 93 ASCII message characters per tweet.
posted by exphysicist345 at 4:40 PM on March 21, 2010
The article on base64 over at Wikipedia has an example showing how every three ASCII characters are encoded as four base64 characters, an inflation of 33% in the message size, so using base64 actually means you can send only 93 ASCII message characters per tweet.
posted by exphysicist345 at 4:40 PM on March 21, 2010
93 is the new 140!
posted by chairface at 7:10 PM on March 21, 2010 [1 favorite]
posted by chairface at 7:10 PM on March 21, 2010 [1 favorite]
odinsdream- good point- the whole message is one, easy to manipulate string that you can shuttle around at will. I noticed the no spaces thing, but just thought it was a coincidence.
posted by gjc at 7:40 PM on March 21, 2010
posted by gjc at 7:40 PM on March 21, 2010
Well in theory you could use the fact that base64 is 8-bit clean to first compress the text message and then base64 encode it, possibly resulting in a net savings for >140 chars. It would be more efficient still to combine compression with an encoding based on the whole unicode code space rather than base64. But that does not appear to be what is going on here nor is it what the client in the above link is doing.
posted by Rhomboid at 8:05 PM on March 21, 2010
posted by Rhomboid at 8:05 PM on March 21, 2010
It's also a good way to send binaries over a text-only channel. You could send small programs, photos, even mp3s etc. Small!
Tho I suppose you could send them split into multiple messages.
posted by blue funk at 1:39 PM on March 22, 2010
Tho I suppose you could send them split into multiple messages.
posted by blue funk at 1:39 PM on March 22, 2010
This thread is closed to new comments.
>>> base64.b64decode('TkJDIE5ld3M6IG11bHRpcGxlIHNvdXJjZXMgc2F5IFN0dXBhayBpcyBhIHllczsgcHJleiBleGVj')
'NBC News: multiple sources say Stupak is a yes; prez exec'
Absurd, but the first is base64-encoded:
'NBC News: multiple sources say Stupak is a yes; prez exec'
The second one is probably also an attempt at base64 encoding, but isn't properly done.
posted by tmcw at 3:50 PM on March 21, 2010