Adam Tornhill - Lisp for the Web

Lisp has kept me fascinated since I hesitantly opened my first pair of parenthesis. To this day, the lessons I took away from Lisp keep informing my decisions in other languages and technologies. Learning Lisp radially changed how I view programming.

But starting with Lisp wasnt a smooth transition. Lisps power comes from a philosophy thats absent in all mainstream languages of today. In a way, learning Lisp was like facing programming as an absolute beginner again.

When it comes to learning, theres nothing more efficient than teaching someone else. Suddenly your cloudy thoughts need to be crystalized, put in context and communicated. In short, you need to truly understand your topic. Following that approach I set out to write an article on Lisp programming. I chose the web as my domain since my preliminary goal was to understand how a 50 year old language could provide just the right tools for an environment that wasnt even conceived back then. If Lisp indeed made it, there had to be something timeless over it. Some inherent quality that could give the programmers of today unequalled power and flexibility.

Writing the original Lisp for the Web article back in 2008 proved to be a rewarding experience. Not only did I succeed on this step in my learning journey; the feedback I got from fellow programmers around the world was overwhelming. I never thought Id reach such a wide audience. In a way Lisp for the Web seemed to provide a natural next step for people that read Paul Grahams essays and were now looking for practical material to take their first step into the exciting (and sometimes frightening) world of Lisp.

Since that time back in 2008 there has been several additions to the Common Lisp ecosystem. Quicklisp lowered the hurdle by providing a simple way to get started, making all powerful libraries easy to install. New books turned up. In particular Land of Lisp stands as an excellent example on a pedagogical introduction to the language. The web development community continued to evolve too. Just to mention a few examples, Edi Weitz made several improvements to his Hunchentoot web-server, NoSQL backends gained popularity, and Matthew Snyder expanded on my work with his excellent Lisp for the Web, part II article.

To this day I keep getting questions on Lisp for the Web. Part of the reason is that I left the original code to rot. I moved on to pursuit other adventures and never really looked back. The result is that the code is now in a sad state where it isnt compatible with newer versions of Hunchentoot. Thats not where I wanted it to end. I think the Common Lisp community deserves up-to-date material; I dont want anyone come to my tutorial with the intentions of learning Lisp and leave out of frustration as the examples dont compile.

This book is my attempt to restore Lisp for the Web to its former glory. Besides bringing the code up-to-date, Ive also taken this opportunity to switch to HTML5 and include some new material. This additional material includes:

• coverage of interacting with the popular NoSQL MongoDB,
• a treatment of the MapReduce algorithm in Lisp and how we can execute it on a remote node without ever leaving our Lisp environment, and
• expanded coverage on how to express JavaScript in Lisp and generate it dynamically, and
• finally, I dive deeper into the Common Lisp Object System (CLOS) and show how it allows us to model complex functionality with simple constructs. The example I include is a variant of the design pattern Observer, but in just one line of standard CLOS code rather than building elaborate class hierarchies.

In the end I hope you enjoy this book and find learning Lisp just as rewarding as I did.

All source code is available in a Github repository. The code comes in three versions: one with the prototypical backend were about to develop in the first part of the book, another version that extends the code to use a persistent storage, and finally a module that includes the MapReduce implementation.

If you clone my Github repository youll notice that Ive duplicated several definitions in the different modules. This isnt something I necessarily encourage in production code. Rather my intent is to present an easily digestible, stand-alone snapshot of each version of the application. In the tutorial I take an iterative approach where I grow a live system organically. Often, this includes replacing existing functions with new versions. As such, the duplications are traces of what once was.

Ive organized this book as a tutorial. If you want to follow along you need a Lisp. There are several high-quality alternatives available to you, both open-source as well as commercial implementations. A good starting point is the Getting Started page on the CLiki.

Common Lisp has tons of interesting open-source libraries. Well get to meet some of them in this book. Since the advent of Quicklisp, installing a library is a breeze. You can get a copy of Quicklisp here: http://www.quicklisp.org. Once youve downloaded your copy, just follow the instructions on the Cliki to install Quicklisp.

The original Lisp for the Web article was published by an Adam Petersen. Now it may look as though Ive stolen not only Mr Petersens idea but also the bulk of his writings. Trust me, its OK - I used to be him. I changed my family name in the summer of 2013 as I married my beautiful Jenny.

Adam is a programmer that combines degrees in engineering and psychology. Hes the author of Your Code as a Crime Scene, has written the popular Lisp for the Web tutorial and self-published a book on Patterns in C. Adam also writes open-source software in a variety of programming languages. His other interests include modern history, music and martial arts.

Thanks to Christopher Wellons, Stuart Malcolm and Matthew Malisz for reporting issues in earlier versions of this book.

My amazing wife Jenny Tornhill designed the cover of this book. Thanks Jenny - you rock!