Connotea

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

~ the obituary ~

Everything you saved to Connotea was public unless you went and said otherwise. That was not a slip, it was the entire mechanism. The site could only tell you that four other people had bookmarked the same paper under the same tag if those four people's libraries were visible, so visible was the default and privacy was a checkbox you had to find.

Nature Publishing Group launched it in December 2004, written in Perl and released under the GPL, which is a stranger thing for a journal publisher to do than it sounds. The pitch was a bookmarking site for scientists at a moment when scientists had no social tools at all. You saved a link and tagged it in whatever words you actually used. The tags stayed uncontrolled on purpose, so that the vocabulary of a field could form on its own rather than being handed down by a librarian.

The trick that made people try it was the metadata. Paste a DOI into the form and Connotea went to CrossRef and came back with the authors and the journal already filled in. For recognized publisher sites it did the same thing straight off the URL. Then it would export the lot in RIS so you could pull it into whatever citation manager you had at home, which meant you could build a bibliography on a machine that had nothing installed on it.

I was at university across most of its life and I did not put my reading anywhere near it. A reading list is unfinished work. Publishing yours tells everybody in your department what you are circling before you have anything to show for it, and I was not going to do that at twenty-two for a service whose payoff was that strangers with adjacent interests could find me. The people it did work for were the ones already writing under their own names in public, which was not most of us.

~ the case file ~

~ paste the doi ~

The DOI intake is the piece worth remembering because everything since has copied it without knowing where it came from. A digital object identifier is a permanent string for a paper and CrossRef is the registry that resolves it. In 2004 almost nothing consumer-facing was touching that pipe, and Connotea sat on top of it and turned a fourteen-character code into a formatted reference in one step.

Around that sat the ordinary bookmarking machinery, mostly done well. Links could be private forever or private until a date you set, which is a small feature with a specific user in mind: somebody sitting on a reference until their own paper is out. Every tag and every user had an RSS feed, so you could subscribe to a subject the way you subscribed to a blog. It won the Association of Learned and Professional Society Publishers award for publishing innovation in September 2005 and got shortlisted for an international award that November.

~ what a publisher wanted with it ~

Nature was not trying to sell subscriptions to Connotea, and the source code was open, so the obvious question is what it was for. The plausible answer is position. In 2004 a publisher could see that the reference layer of science was going to move onto the web and could see that whoever owned it would sit between researchers and journals. Building the tool and giving it away is a cheap way to be in that room.

That bet did not hold, and the thing that broke it was not a competitor with a better product. It was that the room filled up. By 2013 there was CiteULike and Mendeley, plus the general-purpose sites everybody was already on, and a publisher-built bookmarking tool had no argument left. Mendeley was bought by Elsevier three months after Connotea closed, which is the same bet placed later and with money.

~ six weeks to get your library out ~

The announcement went up on the Nature blogs on 24 January 2013 and named the cause plainly: growing problems with spam and the outages that came with them. That is the standard ending for anything that lets the public post links with no gatekeeper. The people generating link spam do not care that the audience is immunologists.

The service was retired on 12 March 2013, and an export tool stayed up until 16 April so registered users could pull their bookmarks out. Six weeks and a working export is a decent ending by the standards of this site, better than most of what is buried here. connotea.org has not answered since, and what it had underneath it, the CrossRef lookup, is still running and still resolving every DOI it was handed.

@}-->--

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

« previous  ·  random  ·  next »