José Miguel Hernandez

50 Cubes, Archived 50cubes.com

Director of Activation · 2010–2011 · San Francisco

Someone else's game engine, carried onto seven platforms and off Flash.

50 Cubes was a San Francisco social-game studio building free-to-play games on Facebook, principally Mall World and Fashion Designer, and syndicating them onto every other social platform that would take them.

Two people I had worked with at Bebo, a colleague and an agency partner whose shop had built for Bebo's advertisers, had started a game studio, and they wanted the platform layer built. This was 2010–2011, when Flash still ran browser games and Facebook was the only distribution that mattered until it wasn't. The game engine predated me and belonged to other engineers; my job, which is what the title meant, was everything around it that got the game in front of players and took their money: the integration layer, the viral loops, the payment flows, the translations.

  • PHP
  • JavaScript
  • Flash
  • OpenSocial
  • SVG

Key projects

Selected work from the 50 Cubes years.

  1. One game, seven platforms

    Distribution

    Mall World was built for Facebook, and then had to run everywhere else. I owned the integrations that put it on MySpace, Orkut, hi5 and Plinga, and through the Mentez and Viximo distribution networks. Orkut was the hardest: OpenSocial cached an XML app definition on its side, loaded our JavaScript and CSS from our domain, asked the gadgets API for the list of app-using friends, handed that back to our server over JSONP to be cached, and only then made signed requests for game state. I drew that bootstrap sequence out as a diagram; every platform needed its own version of it, and each one broke differently.

  2. The viral surface

    Growth

    Neighbor requests, gift sends, invites and feed publishes were mine: hand-written FBML and JavaScript against request forms and profile widgets, tuned for the loops that brought players back. Localization was mine too. When a translation key was wrong, I was the person the QA sheet named.

  3. Credits, offers, and selling real clothes

    Monetization

    I built the payment and offer surfaces through the shift onto Facebook Credits in late 2010, alongside third-party offer walls and in-game ad modals. The most interesting one didn't ship as a business: a tie-up with the retailer Moxsie that put a "buy on Moxsie" button next to virtual items, so a look assembled in a fashion game mapped to clothes you could actually buy. We took their product catalogue, about a thousand items, and built the panel that put it next to the virtual ones.

  4. Getting a Flash game onto the mobile browser

    Runtime

    By 2011 Flash had no future on phones and the studio's entire art pipeline was SWF. Adobe Edge was the product on offer for getting character animation into the browser; we passed on it and built our own. I wrote a batch converter that took the original Flash assets, about fourteen thousand of them, and emitted SVG and PNG, then built the armature that consumed them: an avatar assembled from positioned parts and dressed from the converted catalogue, in a browser runtime targeting iPhone. It took a lot of trial and error, and the cost of getting it wrong was the whole catalogue.

    Sources50 Cubes, archived

The last year was spent trying to get off Flash entirely. It was also where I first had someone reporting to me, a junior engineer named Felix. Managing turned out to be a different job from the one I already knew how to do.

Related context