Mystery Seeker

» see mysteryseeker.com as it lived, on the wayback machine

~ the obit ~

Mystery Seeker was a search box that did not search for you. You typed a query, you hit the button, and what came back was the Google results page for whatever the previous visitor had typed. Your own query went into the queue for the next person. The site's own line for it was that you get what the person before you searched for.

It launched on 2 October 2009 as Mystery Google and ran under that name for seven weeks before becoming Mystery Seeker on 30 November, a rename with an obvious cause given whose results it was serving and whose trademark it had put in the title. The press found it almost immediately. The Telegraph wrote it up on 13 October, HuffPost on the 12th, Mashable in the same stretch, mostly framing it as the "I'm Feeling Lucky" button taken somewhere stranger.

Mechanically it was close to nothing: a front end holding a queue of one, with Google doing the actual retrieval underneath. The whole thing could have been a weekend build, and by every account it was never anything more elaborate than that.

~ the rap sheet ~

~ the missions ~

Within weeks of launch the users had found the real feature. If whatever you type gets shown to the next stranger who arrives, then the search box doubles as a one-line message to a person you will never meet.

So people started coordinating. They called them missions: a group would agree on a phrase and everyone would submit it over and over, so that for a stretch of time anybody who wandered in got the same deliberate sentence staring back instead of a stranger's hunt for a lasagna recipe. Greetings, jokes, pleas, the occasional attempt to organize a meeting point. None of this was designed. It is a messaging system that fell out of a search engine's output because the output happened to be pointed at a stranger.

~ the day the PHP stopped ~

The site just kept running for years, unmaintained and unchanged, serving a page of about 2,320 bytes month after month. The archive has it intact on 2 May 2016, title still reading "MysterySeeker.com: What will you search for?" and the footer still reading Copyright Mystery Seeker, 2009.

Twenty days later it was 1,487 bytes and it was broken in a very specific way. The page being served was no longer a page. It was the raw text of the application's front controller, unexecuted, with the dispatch call and a Configure::read conditional sitting in the open where the HTML used to be. Somebody had moved a host or let a plan lapse, and the new server handed out PHP files instead of running them, so the last thing Mystery Seeker ever published was its own source code.

Nobody fixed it. By May 2017 the domain was answering with a redirect, and by that September it was a 295-byte shell that said "Loading..." and never loaded anything. English-language web toys of this kind rarely reached Turkey on their own, and this one had no reason to: the joke only lands if there is a steady stream of strangers on the other side of your query, and the stream was always going to be American. What stayed with me about it afterward is the shape of the idea rather than the site. Everyone who used it was typing to a person they would never identify, and most of them knew it.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »