Trovebox

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

~ the obituary ~

Jaisen Mathai put OpenPhoto on Kickstarter in late 2011 with a premise that sounds obvious now and was contrarian then: the company that displays your photos should not also be the company holding them. Every photo service of that era worked the same way, in that you uploaded and they kept, and the keeping was the business model. OpenPhoto separated the two jobs. It gave you an interface, an API and apps, and the actual bytes went wherever you told them to go.

Where you told them was usually Dropbox, Box, Amazon S3, or Dreamhost's DreamObjects, and the point of the list was that you were the account holder on all of those. If OpenPhoto disappeared your pictures were exactly where you left them, in storage you were already paying for. The server code was Apache-licensed and on GitHub from the start, so you could skip the hosted service entirely and run the whole thing yourself. Patrick Santana came on as co-founder, and the Shuttleworth Foundation, which funds open-knowledge work rather than returns, put money behind it.

The rename to Trovebox landed in January 2013, partly because OpenPhoto described the license rather than the product. Then in July 2014 Western Digital bought the company. The logic was legible: Western Digital sells My Cloud, a hard drive that sits in your house and pretends to be a cloud service, and Trovebox was a photo layer designed from first principles to run on storage the user owned. On paper it was the rare acquisition where the acquirer actually wanted the thing.

~ the case file ~

~ the escape hatch ~

The shutdown notice went out on 16 January 2015 with a deadline of 15 March to pull an archive of your photos, and the hosted service stopped on 31 March. Two and a half months is generous by the standards of this site, but the export window was close to beside the point here, because most users' photos had never been inside Trovebox to begin with. They were in a bucket with the user's own name on the invoice. The thing being shut down was a viewer.

And the viewer survived too, in the only form that counts. The frontend, the server, the iOS client and the Android client were all still public and still Apache-licensed on the day the service died, so running your own instance stayed possible. At least one IndieHoster picked it up and offered to keep hosting it for people who did not want to administer a server. A product that is architecturally hard to kill is a rare object, and this is what one looks like when the company behind it goes away.

~ the domain that points at its own obituary ~

trovebox.com answers 200 today. It is not a squatter and it is not a parking page. Somebody configured the domain to redirect to the Trovebox article on Wikipedia, so the official homepage of the product is now a third-party encyclopedia entry about how the product ended. I have probed a lot of dead domains for this site and I have not seen another one do that.

~ why i did not run it ~

This one lands close to my territory and I still missed it. Self-hosted, open source, your own S3 bucket, an API you could script against: that is engineer catnip and I was a working engineer through its entire lifespan. What I was not doing in 2013 was administering a photo server. My pictures went where everyone's pictures went, into whatever was already on the phone, and the honest reason is that convenience beats principle for almost everybody including the people who can read the source. Trovebox needed users who would accept a configuration step in exchange for ownership, and there were never enough of us to matter.

@}-->--

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

« previous  ·  random  ·  next »