Slashdot Mirror
International Longest Tweet Contest Seeks Entries
An anonymous reader writes "The 1st International Longest Tweet Contest is open for submissions until April 12. It looks to be a take-off of the famous Obfuscated C Contest. So far the record is 4.2 kilobits encoded per tweet, based on exploiting the fact that Twitter actually passes the full 31 bits of ISO 10646 (the international standard that Unicode is based on), not the roughly 20.08 bits/character of Unicode itself."
140 characters should be enough for anybody!
Huh, that's weird... I still don't have any use for twitter whatsoever.
I mean, I guess I could update the entire world every time I eat something or run an errand... but to be honest, I can't see why anyone who doesn't already know would care, and if they did I think I'd be a bit creeped out by it.
Guess I'm just crazy.
Try not to take me more seriously than I take myself.
Yay. I'll be sure to watch it carefully.
Deleted
first 99999999999999999999 999999999999999999999999999999999999 99999999999999999999999999999999 9999999999999999999 99999999999999 digits of pi
spaces added to evade character filter
Or take a sha256sum and a md5sum or something and make the remote end brute force it (assuming the remote end has a nice quantum computer or something)
perl -le'{print;redo}'
Mine's 140 characters! What do I win?
I think they want you to develop a competing algorithm. Reading the rules, it's "arbitrary binary data" they want to send. Interestingly, a zipped file is one form of "arbitrary binary data", so maybe they'll use that as a test case. If so, your algorithm better be 100% lossless.
<Complete your profile by adding a signature!>
have been doing it for alot longer than us. Often around here outside it sounds like a DDoS.
...but did you know it has actually been implemented?
http://www.rfc-editor.org/rfc/rfc1149.txt
http://www.faqs.org/rfcs/rfc2549.html
http://news.bbc.co.uk/2/hi/africa/8248056.stm
666 exabytes of "6"
about the humongous amounts of information - in bytes - processed every day in tweets. Or email, for that sake. What *happens* to all that information ? All of it together, if backed up, would almost perfectly document our times to posterior generations. Even such contests would, really. Ever thought about that ?
Religous speak to God. Insane are spoken to by God. When all shut up, one can finally hear Shostakovich in peace
I remember a DOS times joke of creating an archive which would expand into a humongous file full of zeroes, 100% lossless btw :) What do I win?
they take up multiple bytes per characters per byte and can encode very complex meanings.
Maybe your life is borring? People I follow usually do more interesting things than just eat and run errands.
I very much doubt that...
Since the decompressor is not part of the tweet data then then this is very simple:
To send: Visit a website post your "message", it get's saved on a server somewhere. The URL is then tweeted to the recipient.
To Receive: Visit the URL and voila there is your message.
No size limits.
Kevin
Back in my day it was about the longest penis. Nowdays its about the longest tweet. 'Nuff said.
C|N>K
Right now, they aren't compressing at all, apparently. 140 characters * 31 bits per character is 4,340 bits, so if they are only getting 4200 bits, they're doing slightly worse than completely uncompressed data here. I'm guessing they have a signature on the front end that loses a few bits.
The question is whether they are using random data or data from real-world sources. If the latter, you might be able to construct a compressor that picks the best of a few hundred compression schemes, then pick the one that got the most data in within 139 characters, and use the 140th 31 bits to encode which scheme you used. If you get lucky and the data happens to be one of the types you're expecting (e.g. uncompressed image data), then you could easily get substantial lossless compression (2:1 should be a piece of cake).
On the other hand, if the source data is truly random (or nearly so, e.g. bytes from the middle of a bzip2 file), then the winner will be equally random, so there's no point in wasting your time being inventive. Instead, just rig the competition by creating a billion Twitter accounts and using the username as additional data bytes.
And, of course, if the person doing the compression can pick the source data, then I can tweet the contents of the Library of Congress with one URL....
Check out my sci-fi/humor trilogy at PatriotsBooks.
Wouldn't the most efficient form of encryption be some kind of steganography? I mean, maybe you wouldn't want to use wikipedia as a steganotext, but receiving a tweet means having internet access, right? If all you have to do is point to a website already containing the unique information you want to send, the competition becomes pretty meaningless, right?
How is it possible that Unicode is 20.08 bits per character? What is that extra .08 bits used for?
For some reason I feel like deleting any post containing the letter-combination "tweet".
Oops.. Dang, then I have to delete this post... Aaargh...
--- Reality doesn't care about your opinions, it happens anyway and if you are in the way you'll get squished.
Isn't that like entering a pointer as your submission in the 'Most Information in 32 Bits' contest?
WTF u twt 2 lng!!
was my first thought of twitter compression.
Paying taxes to buy civilization is like paying a hooker to buy love.
No one has yet found a joke using TL;DR ?
Nope, compression is not part of the contest as the rules state "arbitary binary data" which could be ANY data. Therefore a compression algoritm cannot be used. A compression function basically maps input that are more likely to smaller output at the expense of mapping input that are less likely to larger output. So if your function performs worse than 100% for SOME arbitary binary data, that data breaks the size limit. So the contest is really lame and not related to information technology at all. It's all about finding additional windows to throw data trough. Some "cool suggestions" so far is using your friends list, or creating lots of accounts to transmit data through your username....
They want a system that can store all possible n-bit messages for the largest value of n possible. Your system is bounded by the number of possible urls, so it won't allow you so store, say, all possible 10000-bit messages.
far as I knew, they were 1s & 0s...
well, you write compressor and decompressor so you actually can pass around a url to some data and fetch that.
Good luck trying that. There's tons of spam bots on it that just clog up searches. Nothing useful turns up anyway. It's totally useless. Where do you people find weather emergency alerts and drug cartel movements and stuff anyway? Did you have to sift through "get #viagara #boner http://is.gd/"?
Actually, Twitter is not Unicode-safe.
What happens is you can post a Tweet with astral-plane glyphs and it all appears to work fine, but mysteriously --- a week or so later --- the astral-plane glyphs just vanish. (I don't know if this happens to basic-plane glyphs; I haven't tested it.) I suspect what's happening is that they have different short-term and long-term storage systems, and the long-term systems don't handle Unicode properly.
For example, see this message. That one lasted for about two weeks before the last word vanished. I should probably go hunting for a bug report form...
http://www.penny-arcade.com/comic/2008/4/23/
Most human behaviour can be explained in terms of identity.