ApexKB

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

~ the last words ~

The last release of this thing is dated 15 November 2010. Version 2.0.1.9, GPLv2, posted to SourceForge like everything else it ever shipped. Nothing came after it. The project had been publicly announced on 29 September 2008 under the name Jumper, reached its first stable build on 26 March 2009, and ran out of road about eighteen months later.

What it was: a PHP and JavaScript web application you installed on your own Apache or IIS box, backed by MySQL unless you pointed it at DB2 or Oracle instead. What it did: let a group of people inside an organization tag things, and then search across everything they had tagged, no matter which system the thing actually lived in. The pitch was crowdsourcing your own colleagues. Everyone bookmarks what they know about, the tags pile up, and the pile becomes searchable.

It came out of work Steve Perry had done for Sun Microsystems' Jini Community, which tells you the shape of the problem it was built for: a lot of scattered documentation that nobody could find twice. This is office software for a specific kind of office, and it is not a category I ever touched. I write code in Ankara for a living and I have still never worked anywhere that installed an enterprise bookmarking server.

~ the record ~

~ what it actually did ~

The trick was that a bookmark did not have to be a URL. You could point a tag at a file on a share or at a row inside some other database entirely, and the tag became the thing that was searchable rather than the content. Enough people tagging enough things and you end up with a faceted classification of your company that nobody sat down and designed. That was the theory, and it is a decent one. It also depends completely on people bothering to tag, which is where this whole product category went to die.

Jumper Networks sold commercial support around the free script, which is the standard open-core arrangement of the period. The free version was never crippled. You could run the whole thing without paying anyone, which is why the install base, whatever it was, left no trace when the company behind it went.

~ the two names ~

The project is filed under ApexKB and almost nobody calls it that, because the rename happened after the software had already stopped moving. Jumper is the name on every release, and the name under the frog. The logo was a blue tree frog sitting above the word JUMPER in widely spaced capitals, which is about as much visual identity as a GPL script ever gets.

Both doors are closed now and they are closed in different ways. jumpernetworks.com answers with a 114-byte parking stub, the same one a few thousand other expired domains serve. apexkb.com answers 403 with a Cloudflare page reading "DNS points to prohibited IP," which is what happens when a domain record outlives the machine it pointed at. The code itself is fine. It is sitting on SourceForge under sourceforge.net/projects/jumper, downloadable today, waiting for a server that is not coming.

@}-->--

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

« previous  ·  random  ·  next »