Lisp is used to build applications in which the underlying data is complex and highly dynamic, and where symbolic manipulation is a key feature of the application. Lisp is also used in situations where an application must be readily updated and customized.
You should be aware that Common Lisp is unlikely to become a key component of any commercial operating system in the forseeable future. It is very difficult to merge Lisp storage management techniques with code compiled for traditional pointer-centric languages. Therefore, Lisp must continue to provide its own operating environment. As a Lisp programmer, you should set your expectations accordingly. Once you become facile with Lisp, you'll tend to use Lisp as a workbench and proving ground for algorithmic ideas and programming utilities both large and small. Just don't expect to write and sell a small, simple program in Lisp; the size and cost of the runtime system that you must distribute with your program may be prohibitive.
On the other hand, Lisp excels as a framework for developing large, complex applications. Once past a certain size threshold (for program or data), the fixed overhead of a Lisp environment becomes inconsequential. Also, that rich development environment makes it possible to develop large applications with fewer programmers since much of the algorithmic framework is already provided by Common Lisp's built-in data types and functions. Knowledgeable Lisp programmers achieve even greater productivity by using Common Lisp to write application specific programming tools and language constructs.
The message I want you to take away is that learning and using Lisp is a process. Unless the computer software industry undergoes an unexpected (and unlikely) change, having read this book won't get you a job. However, I'm hoping that what you've learned here will forever color your perceptions as to what is possible, and suggest both techniques and ways of thinking that will be helpful to you in building those "small" applications like word processors, spreadsheets, and OLTP systems.
Meanwhile, I'd like to show you some of the more visible applications that people have developed using Lisp.
More programmers are familiar with Emacs than with any other Lisp application. Richard Stallman conceived Emacs as an extensible editor. He wrote his own Lisp interpreter (not a compiler) specifically for the tasks used in editing text - the low-level text manipulation functions are built-in functions of Emacs Lisp. Over the decades, Emacs has grown to accomodate windowing systems and has accumulated a vast library of code to support programming, writing and personal communications.
Gensym wrote their G2 real-time expert system in Lisp, and later (at greater cost and effort) ported it to C to meet customer expectations. Gensym was able to create a real-time system in Lisp through careful attention to memory allocation; they eliminated unpredictable garbage-collection pauses by simply not generating garbage. (See Chapter 29.)
AutoCAD is a large-scale Computer Aided Design application. A complex design can contain thousands or millions of parts having complex hierarchical and semantic relationships. Like many application developers who face similar problems, the AutoCAD developers leaned heavily on Lisp for their tools and techniques.
Igor Engraver, written by a small team of programmers at Noteheads.com, is an editing and publishing system for musical scores. They've integrated a direct-manipulation graphical interface with a rule-based system that automatically adjusts the layout to conform to complicated rules to meet the expectations of musicians and publishers alike. If that's not enough, Engraver can also play your scores using MIDI instruments. Engraver is targeted to professional composers and musicians, who are encouraged to upload their scores to Noteheads' online e-commerce system for purchase by other musicians. Still want more? Engraver has the slickest software update mechanism you'll find anywhere: select the Check For Patches menu item and Engraver will connect to the Noteheads server, download any patches it needs, and upgrade itself all in a few tens of seconds, without interrupting work in progress.
Yahoo Store is one of the best high-profile Lisp success stories of the past few years. Paul Graham and his team, working out of an attic loft, built and operated a server-side e-commerce site builder for hundreds of customers using Common Lisp and generic Linux servers. Paul's venture was so successful that it drew the attention of Yahoo, who saw it as a better tool for their online stores. Paul sold his company to Yahoo for $49 million. An interesting aside is that Paul hired a couple dozen extra programmers during Yahoo's due diligence investigations, since "no one would believe that three guys in a loft" could have done what Paul's team accomplished with the help of Lisp. Perception is everything...
Each vendor has its own list of success stories. See Chapter 33 for some starting points.