Scroogle

» see scroogle.org as it lived, on the wayback machine

~ the eulogy ~

the objection scroogle was built on took me about a decade to feel. through the 2000s i typed what i wanted into google.com.tr on a machine carrying a cookie that had been sitting there since whenever i last reinstalled, and the idea that the searches were themselves the product had not arrived yet, at least not with me. daniel brandt had it in 2003.

the mechanism fits in a sentence. scroogle asked google on your behalf. you typed into brandt's box and his server ran the query, and back came google's results on a stripped page with the sponsored links removed. google saw brandt's server. it did not see you.

it was not a curiosity either. by the end it was handling something like three hundred and fifty thousand searches a day and sitting inside the top four thousand sites on the web, higher than that in the united states and britain. there was an ssl version for people who wanted the connection encrypted on top, and browser plugins so it could take over the search box.

brandt was not a startup. he was one man with a grievance, running google watch and namebase off the same stubbornness, with wikipedia watch added later. all of scroogle rested on google tolerating a scraper that broke its terms of service, and for years google did, quietly whitelisting it. that arrangement is the part i find hardest to picture now.

~ the dossier ~

~ what it actually did ~

worth making a distinction here, because scroogle gets remembered as a private search engine and it was never one. it had no index of its own. every result it showed you came from google, fetched seconds earlier. what it removed was the link between the query and the person typing it. google got the words and brandt's server address, and nothing else.

the plain results page had a side effect nobody planned. with the ads and the sponsored blocks gone, what was left was ten links on white, which is what google had looked like in 2001. a good number of people used scroogle for that alone. it is the earliest version of an argument that has not stopped since.

~ the december attacks ~

the trouble started at the technical layer. from 2010 google began throttling requests coming from scroogle's servers, which for a service that is nothing but requests to google is a slow strangling. brandt later put a figure on it. even with everything else going fine, he reckoned scroogle had about six months left on the throttling alone.

in december 2011 something else arrived. a sustained denial of service attack from a source nobody ever identified, running around the clock, costing him seven servers over the following weeks as one provider after another dropped him. in february 2012 google's blocking tightened again on top of that. the site put up an outage notice pointing at google. a week later there was no site.

~ what came after ~

brandt did not simply close scroogle. he took every domain he owned out of dns and said why: he wanted whoever was attacking him to understand there were no servers left to trash, so that his old providers would be left alone. his line to searchengineland on 21 february 2012 was that scroogle.org was gone forever, and it was.

the role did not disappear with him. duckduckgo had opened in 2008, and startpage was doing a version of the same proxy idea with google's cooperation rather than against it, which is the difference that let it live. what died with scroogle was the amateur version, one man with seven servers standing between a search box and the largest advertising company in the world. that was never going to hold. it held for nine years.

@}-->--

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

« previous  ·  random  ·  next »