Foreword to the Previous Edition

From the beginning, the Rails framework turned web development on its head with the insight that the vast majority of time spent on projects amounted to meaningless sit-ups. Instead of having the time to think through your domain-specific code, you’d spend the first few weeks of a project deciding meaningless details. By making decisions for you, Rails frees you to kick off your project with a bang, getting a working prototype out the door quickly. This makes it possible to build an application with some meat on its bones in a few weekends, making Rails the web framework of choice for people with a great idea and a full-time job.

Rails makes some simple decisions for you, like what to name your controller actions and how to organize your directories. It also gets pretty aggressive and sets development-friendly defaults for the database and caching layer you’ll use, making it easy to change to more production-friendly options once you’re ready to deploy.

By getting so aggressive, Rails makes it easy to put at least a few real users in front of your application within days, enabling you to start gathering the requirements from your users immediately rather than spending months architecting a perfect solution only to learn that your users use the application differently than you expected.

The Rails team built the Rails project itself according to very similar goals. Don’t try to overthink the needs of your users. Get something out there that works and improve it based on actual usage patterns. By all accounts, this strategy has been a smashing success, and with the blessing of the Rails core team, the Rails community leveraged the dynamism of Ruby to fill in the gaps in plugins. Without taking a close look at Rails, you might think that Rails’ rapid prototyping powers are limited to the 15-minute blog demo but that you’d fall off a cliff when writing a real app. This has never been true. In fact, in Rails 2.1, 2.2, and 2.3, the Rails team looked closely at common usage patterns reflected in very popular plugins, adding features that would further reduce the number of sit-ups needed to start real-life applications.

By the release of Rails 2.3, the Rails ecosystem had thousands of plugins, and applications like Twitter started to push the boundaries of the Rails defaults. Increasingly, you might build your next Rails application using a nonrelational database or deploy it inside a Java infrastructure using JRuby. It was time to take the tight integration of the Rails stack to the next level.

Over the course of 20 months, starting in January 2008, we looked at a wide range of plugins, spoke with the architects of some of the most popular Rails applications, and changed the way the Rails internals thought about its defaults.

Rather than starting from scratch, trying to build a generic data layer for Rails, we took on the challenge of making it easy to give any ORM the same tight level of integration with the rest of the framework as Active Record. We accepted no compromises, taking the time to write the tight Active Record integration using the same APIs that we now expose for other ORMs. This covers the obvious, such as making it possible to generate a scaffold using DataMapper or Mongoid. It also covers the less obvious, such as giving alternative ORMs the same ability to include the amount of time spent in the model layer in the controller’s log output.

We brought this philosophy to every area of Rails 3: flexibility without compromise. By looking at the ways that an estimated million developers use Rails, we could hone in on the needs of real developers and plugin authors, significantly improving the overall architecture of Rails based on real user feedback.

Because the Rails 3 internals are such a departure from what’s come before, developers building long-lived applications and plugin developers need a resource that comprehensively covers the philosophy of the new version of the framework. The Rails™ 3 Way is a comprehensive resource that digs into the new features in Rails 3 and perhaps, more important, the rationale behind them.

—Yehuda Katz
Rails Core