Koders

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

~ the goodbye ~

what i wanted in 2005 was not a search engine for pages. it was a search engine for other people's code, because the answer to nearly everything an engineering student runs into is already written down inside somebody's project and the entire difficulty is finding which one. koders is the product that arrived there first. google is the one my memory kept.

darren rush built it and ran it as koders, inc. out of santa monica. the public engine opened in may 2005 with roughly a hundred and ninety million lines indexed off the open repositories of the day, sourceforge and apache among them. you searched by keyword and narrowed by programming language, and what came back were functions and classes rather than pages written about functions and classes. it raised a small round in may 2006, around seven hundred and sixty thousand dollars, with the founders fund on the cap table.

black duck software, a boston company whose business was telling corporations which open source licenses were sitting inside their shipped products, bought the assets in april 2008 for a sum nobody published. under black duck the index went past two point four billion lines by september 2009, and microsoft's codeplex projects were folded into it along the way.

whether the tab open on my machine in those years said koders or said google is not something i can honestly reconstruct now. what i can say is that searching code was a separate errand you went somewhere specific to run, and it stayed separate for about a decade before the code hosts absorbed it back.

~ the chronicle ~

~ what it actually did ~

a web search engine reads code as text. koders parsed it. it knew a line was a method definition rather than a comment, so searching a function name returned the place it was declared instead of every forum thread arguing about it. results came back with the license attached, which was the detail that mattered if you were about to paste something into work you were being paid for.

there was a business sitting behind the free site. the paid version ran inside a company firewall and searched the code that company already owned, on the reasonable theory that a large engineering organization has written the thing you need at least twice and can find neither copy. koders shipped plugins for eclipse and visual studio so the search lived where the work was rather than in another browser tab.

~ the year google showed up ~

google code search opened in october 2006, sixteen months behind koders, with regular expression support and google's name on the door. it took the attention immediately. it also died first, switched off in january 2012 in one of google's clearing rounds, which left koders standing as the survivor of a fight it had already lost on recognition years earlier.

by then black duck owned two of these. it had bought ohloh in 2010 and was running ohloh code alongside koders with no good reason to keep both. on 25 october 2012 the two were merged into a single site called open hub, and koders.com stopped serving its own pages inside a week. open hub's code search was itself discontinued in 2016. the domain now lands on synopsys, which bought black duck in 2017, and there is nothing at the end of it that would tell you what used to be there.

@}-->--

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

« previous  ·  random  ·  next »