Yahoo! SearchMonkey

» see developer.yahoo.com/searchmonkey as it lived, on the wayback machine

~ the obit ~

By 2008 nobody around me ran a search that was not Google. Yahoo had lost Turkey years earlier and was not coming back, so the proposition here, spend an afternoon adding structured markup to your pages so they look better in Yahoo results, was work attached to an audience that had already left. SearchMonkey was a clever thing happening in a room I had stopped walking into.

The idea itself was correct and it was early. Yahoo opened SearchMonkey in May 2008 and named it after Greasemonkey, the Firefox extension that let you rewrite other people's web pages in your own browser. Same instinct, aimed one layer up: instead of the browser, the search result. If you owned a site, you could take control of how your listing rendered on the results page.

You did it by putting data on your pages in a format the crawler understood. Microformats, RDFa, eRDF, or a feed submitted straight to Yahoo. Yahoo's crawlers picked that up and made it available to anyone writing an app, and the app then decided what the result looked like. Two kinds existed: a Result App that replaced the listing entirely, and an Infobar that hung underneath it.

~ the rap sheet ~

~ how it landed ~

Developers liked it more than users noticed it. The presentation half ran on a stripped-down PHP hosted on Yahoo's own servers, and the data half could scrape a page with XSLT or chain calls out to arbitrary web services and stitch the answers together before the search page finished rendering. That is a lot of machinery to hand to strangers, and Yahoo handed it over in public.

The demo apps were the enjoyable part. Search Amazon for a specific DVD and the result came back enriched. Search YouTube and the video was inline in the listing, which meant a well-chosen query could Rickroll you from inside a Yahoo search result in 2008. ReadWriteWeb put it in the top ten semantic web products of the year, back when the semantic web still had a top ten.

~ the actual end ~

In July 2009 Yahoo agreed to hand its search back end to Microsoft. The transition ran through 2010, and everything Yahoo had built on top of its own index became something with no index under it. SearchMonkey went off on 1 October 2010, with the developer tools and the application gallery pulled the same day. Site Explorer and the free tier of BOSS went in the same clear-out.

Nobody at Yahoo pretended otherwise. Computerworld's write-up in August 2010 had already called it a casualty of the deal, which is the plainest possible description. The product had no defect. Its host stopped operating a search engine, and a feature of a search engine cannot outlive the search engine.

~ the next chapter ~

Google announced rich snippets in May 2009, a year after SearchMonkey opened, and by 2011 the major search engines had agreed on schema.org as the shared vocabulary for exactly the thing Yahoo had been asking site owners to do. Every structured-data result you see now, the recipe with its rating stars or the event with its date already filled in, descends from the same premise. Yahoo got there first, built the developer platform for it, then sold the engine it all depended on.

@}-->--

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

« previous  ·  random  ·  next »