Pubget

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

~ the obit ~

Ramy Arnaout was a clinical pathologist at Beth Israel Deaconess in Boston, and like everyone else doing that job he spent part of every week trying to get hold of papers. The subscriptions existed. His institution had paid for them. Getting from a citation to the actual document still meant a chain of redirects through a publisher's site, a proxy login, a landing page, an abstract and finally a download link, repeated for every paper and different on every publisher. He started Pubget in 2007 to shorten that.

The product was a search box that returned PDFs. You typed what you were looking for and Pubget resolved it against your institution's holdings and put the document on screen, which sounds like a small thing until you count how many steps it removed. The service was free for non-profit institutions and their libraries, and at its height it reached full-text content across 450 libraries around the world.

It stayed inside the licensing system. Pubget did not route around paywalls; where a paper was restricted and the institution had not paid, it showed the abstract and stopped. The whole pitch to libraries was that it made the access they had already bought easier to use, which is also why libraries were willing to point their users at it. The company moved from Cambridge to Boston's Innovation District in 2011.

In January 2012 Copyright Clearance Center acquired it, and Pubget Corp became a wholly owned subsidiary developing cloud-based search for scientists. Nearly six years later CCC turned it off.

~ the rap sheet ~

~ the paperwall ~

The problem Pubget solved never went away, so it is worth setting out. Scientific literature is sold by publishers to institutions, and the person reading it is a third party to that transaction who experiences it as a series of login walls that do not know about each other. A researcher at a well-funded university has legitimate access to most of what they need and still cannot get at it in under a minute.

Four years into Pubget's life, Sci-Hub launched and solved the same problem by ignoring the licensing entirely. It is now among the most heavily used research tools anywhere. Pubget was the version that respected every subscription boundary and asked the libraries for permission first. Sci-Hub outlived it by years.

~ 1 october 2017 ~

The archive puts the ending inside two weeks on its own: pubget.com was serving its normal six-kilobyte homepage on 29 September 2017 and 1,190 bytes of shutdown notice on 12 October. The notice says the date outright, which saves anyone the guesswork. "Pubget.com shut down on October 1, 2017 after ten years of helping users find papers fast." It then points readers at copyright.com and thanks them for being a loyal customer, over a Copyright Clearance Center copyright line.

The number in the pitch was 450 libraries, and that number is also the limit of the product. Pubget worked by assuming the subscriptions were already bought, which made it excellent for a well-funded institution and irrelevant everywhere else. Studying engineering in Turkey through those years, the question was never which tool resolved a PDF fastest; it was whether the license existed at all, and nothing about that was visible from a search box. Ten years of helping users find papers fast is an accurate epitaph. It is also a quiet admission about which half of the problem the company was ever able to reach.

@}-->--

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

« previous  ·  random  ·  next »