Dapper

» see open.dapper.net as it lived, on the wayback machine

~ the eulogy ~

Dapper called itself The Data Mapper, which undersells it. You gave it the address of a web page, and it loaded the page in front of you so you could drag a selection box around the part you actually wanted. Dapper worked out the pattern from that one example and applied it to every other page on the site. You walked away with a feed, having asked nobody's permission and used nobody's API key.

Eran Shir and Jon Aizen founded it in late 2005 with seed money from Accel, and the public service opened in 2006. The output side was the clever half. The same extracted data could come back as RSS or as a Google Gadget, so a site that published nothing but HTML could be made to behave like one with a developer program. For about four years this was the standard answer to a problem that came up constantly: the information is right there on the screen and there is no way to subscribe to it.

Yahoo bought the company on 5 October 2010 for a reported fifty-five million dollars, and the thing it wanted was the advertising product. The scraper came along with it. Dapper's engineering team became the core of Yahoo's Creative Innovation Center, which built dynamic ad creative, and Eran Shir ran it until he left for Israel in November 2013. The free tool at open.dapper.net was not part of that plan and was not shut down either. It simply kept running, under a logo Yahoo had repainted with "from YAHOO!" underneath the wordmark.

~ the dossier ~

~ the day it stopped ~

There is no announcement to find, so the date comes out of the archive. On 22 July 2015 open.dapper.net was still serving its real homepage at around 3.7 KB, Yahoo analytics tag and all. Five days later, on 27 July, the working endpoints were gone: a request to transform.php, the URL that actually produced a feed, returns 404. From 10 August the homepage itself was a 248-byte page reading "The url of this site has been shutdown. Please update your bookmarks!" above a meta refresh to yahoo.com.

Yahoo had announced a cull six weeks earlier, on 4 June 2015. The coverage named Maps and Pipes at the top and worked down through the developer APIs and the regional verticals. Dapper is not on that list anywhere. It was switched off in the same season by the same company and never made the press release.

~ who was still using it ~

The archive crawled those dead transform.php URLs along with the homepage, and the query strings are still readable, which means the last requests Dapper ever failed are a matter of record. Two of them were pulling news feeds out of Tamil-language newspapers, Daily Thanthi and Dinamani. Further down there is a Romanian design studio's blog and a German printer manufacturer's success-stories page.

That is what a tool like this gets used for, and it is not what the acquisition deck said. These were people outside the English-language web with a site they needed to follow and no feed to follow it with, which was also the ordinary condition of reading anything in Turkish online in those years. When Dapper went, every one of those feeds stopped filling, and nobody was told why.

~ the next chapter ~

Yahoo Pipes closed the same year, so 2015 took both halves of the make-your-own-feed idea within months. Nothing picked it up in the same spirit. The job Dapper did for free is now done by scraping APIs that bill per thousand requests, so what went away was the price rather than the capability.

The logo survives better than the service. Fetch it out of the archive and you get a colored pinwheel beside a lowercase wordmark, with "from YAHOO!" stamped underneath in small type.

@}-->--

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

« previous  ·  random  ·  next »