Dinah Autoresponse: Difference between revisions

From Perplex City Wiki
Jump to navigationJump to search
mNo edit summary
mNo edit summary
Line 1: Line 1:
This text was received after someone dropped an email to the dinah@projectsyzygy.com address (from the [[Marketing Weekly]]). Received an autoresponse back:
This text was received after someone dropped an email to the dinah@projectsyzygy.com address (from the [[Marketing Weekly]] ad):


     On 5 Mar 2004, at 00:29, xxxx wrote:  
     On 5 Mar 2004, at 00:29, xxxx wrote:  

Revision as of 21:40, 25 February 2005

This text was received after someone dropped an email to the dinah@projectsyzygy.com address (from the Marketing Weekly ad):

   On 5 Mar 2004, at 00:29, xxxx wrote: 
   
   I don't think I have anything more to add, except for the requisite  shout-outs to the immersive
   fiction community all over the  web and  around the world (you know who you are). Anyone reading
   this should  know that it's going to get a lot harder and a lot more fun.  You ain't seen nothing yet. 
   
   One last thing - it all begins with AADAA... 
   
   xxxx
   
   
   
   On 5 Mar 2004, at 00:13, xxxx wrote: 
   
   What a cool idea.  I love it.   All we'd need to do is sanitise our email exchange by removing reference to things like xxxx and
   then put it online somewhere.  This could be the last email of our exchange we publish.
   
   Unless of course you follow this one with a witty reply that our  hard-core cryptographer friends would enjoy ;)
   
   xxxx
   
   
   
   On 4 Mar 2004, at 23:58, xxxx wrote: 
   
   Actually, yeah, I think you are probably right. As long as the puzzle   isn't *ridiculously* easy, then people aren't going to
   think worse  of  it (especially if you put 'Easy wasn't it?' as a more difficult  one!).  Simple substitution is a nice way to go. 
   
   Now I'm wondering what sort of code to use for the 'easy wasn't it'   thing...  
   
   It occurred to me this evening that a nice reward for ultra-hardcore   cryptographers would be to see this exchange of emails - 
   it wouldn't   actually reveal anything at all about xxxx, and they'd probably   appreciate the thought that's gone into the ad
   puzzle. Perhaps we  could  encrypt a key in the ad that would unlock these emails on the  web  somewhere. Or maybe not. Just 
   another random idea  
   
   xxxx 
   
   
   
   On 4 Mar 2004, at 23:48, xxxx wrote: 
   
   Fair point. I've been thinking about this a lot this evening. It's a   tricky balance we're trying to strike. 
   
   The code at the moment is laughably easy to crack but that's the   point. I think 99% of people reading the ad will  feel pretty
   pleased with themselves for cracking a Caesar  shift or  simple substitution cipher. The image I have is of people  stumbling
   across our ad while flicking through a stack of job ads  in a coffee  shop. They'll be intrigued by it but unless they have a  
   hook to start  solving it quickly I think they'll give up and  probably won't give it  another thought. 
   
   Alternatively if they immediately see a way in and manage to solve  it  on the back of a napkin within five or ten minutes then 
   I think  they'd  be very likely to check out the site and more importantly  mention it  to other people. 
   
   In hindsight a simple substitution cipher probably makes more sense   than a Caesar shift - particularly if we leave the web
   address  format  as it is. This provides the easy way in to give the average  person a  foothold of four letters. 
   
   It'll be a different matter when the game is live and people are xxxx but in this instance I think a Vigenere cipher would require
   too much effort for the average  person.  Obviously we wouldn't want to recruit anyone for the  full-time team  who'd struggle to
   decode a Vigenere cipher but  this is a great  opportunity to get a lot of people chatting about  the concept and the  site itself.
   If the puzzle is too complex then  we lose all that and  the fact we gain an average of higher quality  submissions probably
   doesn't offset it. 
   
   You're right in that we don't want to look like lightweight puzzle   designers so a way around this might be to hide a much more
   complex   code in the background that acknowledges the very important but tiny minority of ultra smart folks. Perhaps something 
   as simple as "Easy wasn't it?" 
   
   Any thoughts?? 
   
   Cheers 
   
   xxxx
   
   
   
   
   On 4 Mar 2004, at 18:38, xxxx wrote: 
   
   Hi all, 
   
   I think the text of the ad is perfect, but I have some comments on other aspects. 
   
   Firstly, I think a Caesar shift is too easy - I mean, even I could   figure that one out! If people are interested in the ad, then
   they'll  be willing to put at least a few minutes work into it -  and if  they're not interested in the ad, it won't be because 
   the cryptosystem is too hard, it'll be because it just looks like   gibberish. So I don't feel there's anything wrong with 
   knocking it  up  a notch and using, say, a simple substitution cipher. These are   easily broken using frequency analysis or 
   brute force, but it's a   step up from Caesar. 
   
   Another possibility is using a Vigenere Cipher, which is essentially a Caesar cipher done twice, using a keyword to determine the
   shift.  It's much more difficult to brute force without the keyword and very  very difficult for short texts, which  obviously 
   this one is, so we  should make the keyword available. A possible scenario would be to encode the majority of the text in a 
   Vigenere Cipher (which is immune  to frequency analysis) and encode  the keyword (written in Braille or  morse code or baudot) 
   with a  simple substitution cipher or something  similar.
   
   Whichever system is used, it would be a mistake to write out the  web  address in the usual 'http://www....com' format, simply
   because it'd  be very recognisable and thus provide an easy  shortcut to crack the  puzzle (and that would be no fun for  anyone).
   To be totally safe, I  would write it out as 'visit  projectsyzygy dot com' so there isn't  any strange punctuation to  give the
   game away. 
   
   So - if you went for the Vigenere route, here is how I would do it   (and obviously there are many other ways, many of which are
   undoubtedly better): 
   
   1. In a corner or on the border of the ad, the phrase 'You're  always  a day away' is written in morse or braille or baudot.
   Optionally, it  is encoded via a simple substitution cipher, or  maybe just a Caesar  shift. 
   
   2. This phrase is obviously from the song 'Tomorrow' in the movie   'Annie'. Thus the keyword is 'Tomorrow' - it's not hard to
   figure out. 
   
   3. From there, if you recognise that the main text is encoded in   Vigenere, it is trivial to decode it using the keyword. 
   
   You could always highlight some of the letters in the text and use   them as a keyword for a future puzzle on the web - we 
   wouldn't even   need to decide on them beforehand, we could just design the future puzzle around them. 
   
   My reasoning for making the puzzle a bit harder and more  circuituous is because it involves the 'player' a little more and
   thus will give  them a bit more satisfaction on completion and also convince them  that we are not lightweights when it comes 
   to cryptography.  Furthermore, if anyone did want to write a story about it, it's more  interesting this way  
   
   Of course, it is entirely possible that making the puzzle harder  will  put people off, which is clearly not the goal. However,
   it will be cracked by *someone* in short order, and the answer and  instructions  will then be put on the web for all to see.
   Success!   
   
   
   xxxx
   
   
   On 4 Mar 2004, at 16:32, xxxx wrote: 
   
   Here's a first stab at text for the ad that will be going into the press over next week or so. I'm planning to put it into 
   Marketing Week and then play it by ear after that. 
   
   If anyone has any suggestions or improvements then give me a shout asap since we're hoping to get this finished by midday on 
   Monday. Might be nice to hide a deeper clue in there somewhere as well.  Any  suggestions? Perhaps we could very subtly highlight
   some of  the  letters (tiny dot above or below) to spell out a new message.  Or we  could partially hide the 'I can see you'
   Braille text  somewhere  within the design around the edge of the text? Not very imaginative though! 
   
   I currently think that the best plan would be for the first  sentence  to be unencrypted to hook people's interest and then
   encode the rest of the text via a caesar shift (including the web  address  obviously!) 
   With a bit of luck it should get picked up and commented on by the   mainstream press.
   
   Cheers 
   
   xxxx