By 2010 the plugins were losing. Steve Jobs published his "Thoughts on Flash" letter in April, the iPad launched without Flash support, and HTML5 video and canvas became the way to do animation and media on the web. For a studio that had sold a lot of Flash work since 2007, this was the year to retool.
From Flash to HTML5
We moved product animations and interactive presentations to HTML5, CSS3 and JavaScript over the course of the year. Rounded corners, gradients and shadows no longer needed sliced images; @font-face gave us real typography; and HTML5 form inputs and local storage made web apps feel closer to desktop software. Older browsers needed polyfills and graceful fallbacks, so every project carried a progressive-enhancement layer, but the direction was obvious.
Cloud hosting for real projects
2010 was also the year "the cloud" stopped being a buzzword for us. Renting virtual servers by the hour meant a client's web application could start on a small instance and grow with traffic instead of guessing the hardware up front. We moved our first production applications off shared hosting and onto cloud servers, learned about backups, monitoring and deployment the hard way, and started offering cloud hosting as a service line, which it still is.
What clients were asking for
- Web applications rather than websites: ordering, booking, membership and reporting tools built on PHP and MySQL.
- Video on the web, now that HTML5 video and YouTube embeds worked on phones and tablets.
- Mobile-friendly layouts for the iPhone, Android handsets and the new iPad.
The team ended 2010 with a clear split in the work: design and multimedia on one side, a growing custom development practice on the other. The following year, with responsive design, the two would meet in the middle.
Looking ahead
The next step was already decided: responsive design. Maintaining a desktop site and a separate mobile site for every client was doubling the work and halving the quality, and the media-query approach that Ethan Marcotte had described in May 2010 was clearly the answer. We spent the last quarter of the year building our first fluid, single-codebase layouts and retraining the design team to think in proportions rather than pixels. On the application side, the custom PHP work had grown large enough to need structure, and we began evaluating the new generation of PHP frameworks that would eventually lead us to Laravel. Finally, the cloud servers we had just learned to run needed proper monitoring, backups and deployment scripts; the support team spent the winter writing them. 2011 would be the year all of this came together.
