Tapestry Tutorial: Introduction

What is Tapestry?

Welcome to Tapestry!

This is a tutorial for people who will be creating Tapestry 5 applications. It doesn't matter whether you have experience with Tapestry 4 (or Tapestry 3, for that matter) or whether you are completely new to Tapestry. In fact, in some ways, the less you know about web development in general, and Tapestry in particular, the better off you may be ... that much less to unlearn!

You do need to have a reasonable understanding of HTML, a smattering of XML, and a good understanding of basic Java language features, and a few newer things such as Java Annotations.

If you're used to developing web applications using servlets and JSPs, or with Struts, you are simply used to a lot of pain. So much pain, you may not even understand the dire situation you are in! These are environments with no safety net; Struts and the Servlet API has no idea how your application is structured, or how the different pieces fit together. Any URL can be an action and any action can forward to any view (usually a JSP) to provide an HTML response to the web browser. The pain is the unending series of small, yet important, decisions you have to make as a developer (and communicate to the rest of your team). What are the naming conventions for actions, for pages, for attributes stored in the HttpSession or HttpServletRequest?

The traditional approaches thrust something most unwanted in your face: multithreaded coding. Remember back to Object Oriented Programming 101 where an object was defined as a bundle of data and operations on that data? You have to unlearn that lesson as soon as you build a web application, because web applications are multi-threaded. An application server could be handling dozens or hundreds of requests from individual users, each in their own thread, and each sharing the exact same objects. Suddenly, you can't store data inside an object (a servlet or a Struts Action) because whatever data you store for one user will be instantly overwritten by some other user.

Worse, your objects each have one operation: doGet() or doPost().

Meanwhile, most of your day-to-day work involves deciding how to package up some data already inside a particular Java object and squeeze that data into a URL's query parameters, so that you can write more code to convert it back if the user clicks that particular link. And don't forget editing a bunch of XML files to keep the servlet container, or the Struts framework, aware of these decisions.

Just for laughs, remember that you have to rebuild, redeploy and restart your application after virtually any change. Is any of this familiar? Then perhaps you'd appreciate something a little less familiar: Tapestry.

Tapestry uses a very different model: a structured, organized world of pages, and components within pages. Everything has a very specific name (that you provide). Once you know the name of a page, you know the location of the Java class for that page, the location of the template for that page, and the total structure of the page. Tapestry knows all this as well, and can make things just work.

As well see in the following chapters, Tapestry lets you code in terms of your objects. You'll barely see any Tapestry classes, outside of a few Java annotations. If you have information to store, store it as fields of your classes, not inside the HttpServletRequest or HttpSession. If you need some code to execute, its just a simple annotation or method naming convention to get Tapestry to invoke that method, at the right time, with the right data. The methods don't even have to be public!

Tapestry also shields you from the multi-threaded aspects of web application development. Tapestry manages the life-cycles of your page and components objects, reserving particular objects to particular threads so that you never have to think twice about threading issues.

Tapestry began in January 2000, and now represents over seven years of experience: not just my experience, or that of the other Tapestry committers, but the experience of the entire Tapestry community. Tapestry brings to the table all that experience about the best ways to build scalable, maintainable, robust, internationalized (and more recently) Ajax-enabled applications. Tapestry 5 represents a completely new code base designed to simplify the Tapestry coding model while at the same time, extending the power of Tapestry and improving performance.

About Releases and Documentation

If you are reading this documentation online, be aware that the online documentation generally represents the very latest version of the code, termed "the snapshot". The documentation is written as if the next release is available, but may reference version numbers, or even features, that are only available by building the latest Tapestry source.

About the Author

Howard Lewis Ship is the creator of Tapestry, and the Chair of the Tapestry Project Management Committee at Apache. Howard is the Director of Open Source Technology at Formos Software Development. Howard lives in Portland, Oregon with his wife Suzanne, a novelist.


Continue on to Chapter 1: Setting Up Your Environment