Just Like Starting Over
In April 2004, Kory Juul and I had an idea for a production pipeline that didn’t suck. We discussed the pipelines that we’d seen over the years doing film effects and architectural rendering and said to each other, “There must be a better way!” We went off in search of the better way and discovered some cool concepts and came up with a framework for our pipeline. And then, in between all the work, we started coding.
After more than two years of dabbling and thinking and hitting some decent code complete milestones (and learning Ruby On Rails in the process), I’ve decided to kill the existing code base and start over. The more we used it, the more we realized that the existing system just didn’t work quite right.
We realized that the core of the application was about elements. An element being the smallest unit in a facility. Any digital asset such as a Maya file, a texture map, motion capture data, and the like. The core of how the application needed to work led us to think about how an element is born and moves through the system. Everything revolves around the element and what various people in myriad departments do with it. In my first stab, I had thought more about the project and less about the element and the shot.
Also, in the last two years, Rails has come a very long way from a small beta framework to the now-release-candidate-grade-version 1.2. We’ve got Mongrel and Components have gone the way of the Dodo. The framework has moved towards a RESTful approach. It’s delicious.
We’ve also become a bit more experienced in how we think a pipeline should run. Finishing The Sandbox and working on a commercial for Microsoft and some shots for a major studio feature gave us some new insight.
The interface won’t change much, but the back end will change immensely. We’re going RESTful. We’re utilizing Test Driven Development (I hate writing tests, but love them later in development). We’re focusing more on a great Digital Asset Manager than a hacked up project management application.
And it’s not too shabby. In two evenings I have a basic core with working, tested code. I’ve got realistic sprints with defined user stories. And hopefully, by NAB, we’ll have a good working prototype to show off.
like starting over.