A Virtual Machine
Safety of Design
Safety of Implementation
Application and User Level Security
Java and the World Wide Web
Java as a General Application Language
A Java Road Map
The greatest challenges and most exciting opportunities for software developers today lie in harnessing the power of networks. Applications created today, whatever their intended scope or audience, will almost certainly be run on machines linked by a global network of computing resources. The increasing importance of networks is placing new demands on existing tools, and fueling the demand for a rapidly growing list of completely new kinds of applications.
We want software that works--consistently, anywhere, on any platform--and that plays well with other applications. We want dynamic applications that take advantage of a connected world, capable of accessing disparate and distributed information sources. We want truly distributed software that can be extended and upgraded seamlessly. We want intelligent applications--like autonomous agents that can roam the Net for us, ferreting out information and serving as electronic emissaries. We know, to some extent, what we want. So why don't we have it?
The problem has been that the tools for building these applications have fallen short. The requirements of speed and portability have been, for the most part, mutually exclusive, and security has largely been ignored or misunderstood. There are truly portable languages, but they are mostly bulky, interpreted, and slow. These languages are popular as much for their high level functionality as for their portability. And there are fast languages, but they usually provide speed by binding themselves to particular platforms, so they can meet the portability issue only half way. There are even a few recent safe languages, but they are primarily offshoots of the portable languages and suffer from the same problems.
The Java programming language, developed at Sun Microsystems under the guidance of Net luminaries James Gosling and Bill Joy, is designed to be a machine-independent programming language that is both safe enough to traverse networks and powerful enough to replace native executable code. Java addresses the issues raised here and may help us start building the kinds of applications we want.
Right now, most of the enthusiasm for Java stems from its capabilities for building embedded applications for the World Wide Web; these applications are called applets. This book will teach you how to build applets. But there is more to Java than applets, and we'll also try to show you the "more." The book will also show you how to use the tools of Java to accomplish real programming tasks, such as building networked applications and creating functional user interfaces. By the end of the book, you will be able to use these tools to build powerful Java applets and standalone applications.
The seeds of Java were planted in 1990 by Sun Microsystems patriarch and chief researcher, Bill Joy. Since Sun's inception in the early '80s, it has steadily pushed one idea: "The network is the computer." At the time though, Sun was competing in a relatively small workstation market, while Microsoft was beginning its domination of the more mainstream, Intel-based PC world. When Sun missed the boat on the PC revolution, Joy retreated to Aspen, Colorado, to work on advanced research. He was committed to accomplishing complex tasks with simple software, and founded the aptly named Sun Aspen Smallworks.
Of the original members of the small team of programmers assembled in Aspen, James Gosling is the one who will be remembered as the father of Java. Gosling first made a name for himself in the early '80s as the author of Gosling Emacs, the first version of the popular Emacs editor that was written in C and ran under UNIX. Gosling Emacs became popular, but was soon eclipsed by a free version, GNU Emacs, written by Emacs's original designer. By that time, Gosling had moved on to design Sun's NeWS window system, which briefly contended with the X Window System for control of the UNIX graphic user interface (GUI) desktop in 1987. While some people would argue that NeWS was superior to X, NeWS lost out because Sun kept it proprietary and didn't publish source code, while the primary developers of X formed the X Consortium and took the opposite approach.
Designing NeWS taught Gosling the power of integrating an expressive language with a network-aware windowing GUI. It also taught Sun that the Internet programming community will refuse to accept proprietary standards, no matter how good they may be. The seeds of Java's remarkably permissive licensing scheme were sown by NeWS's failure. Gosling brought what he had learned to Bill Joy's nascent Aspen project, and in 1992, work on the project led to the founding of the Sun subsidiary, FirstPerson, Inc. Its mission was to lead Sun into the world of consumer electronics.
The FirstPerson team worked on developing software for information appliances, such as cellular phones and personal digital assistants (PDA). The goal was to enable the transfer of information and real-time applications over cheap infrared and packet-based networks. Memory and bandwidth limitations dictated small and efficient code. The nature of the applications also demanded they be safe and robust. Gosling and his teammates began programming in C++, but they soon found themselves confounded by a language that was too complex, unwieldy, and insecure for the task. They decided to start from scratch, and Gosling began working on something he dubbed "C++ minus minus."
With the floundering of the Apple Newton, it became apparent that the PDA's ship had not yet come in, so Sun shifted FirstPerson's efforts to interactive TV (ITV). The programming language of choice for ITV set-top boxes was the near ancestor of Java, a language called Oak. Even with its elegance and ability to provide safe interactivity, Oak could not salvage the lost cause of ITV. Customers didn't want it, and Sun soon abandoned the concept.
At that time, Joy and Gosling got together to decide on a new strategy for their language. It was 1993, and the explosion of interest in the Internet, and the World Wide Web in particular, presented a new opportunity. Oak was small, robust, architecture independent, and object oriented. As it happens, these are also the requirements for a universal, network-savvy programming language. Sun quickly changed focus, and with a little retooling, Oak became Java.
I don't think it's overdoing it to say that Java has caught on like wildfire. Even before its first official release, while Java was still a nonproduct, nearly every major industry player jumped on the Java bandwagon. Java licensees include Microsoft, Intel, IBM, and virtually all major hardware and software vendors.
As we begin looking at the Java architecture, you'll see that much of what is exciting about Java comes from the self-contained, virtual machine environment in which Java applications run. Java has been carefully designed so that this supporting architecture can be implemented either in software, for existing computer platforms, or in customized hardware, for new kinds of devices. Sun and other industry giants have announced their intentions to produce cheap, fast Java chips, the first of which should be available by the time you read this. Hardware implementations of Java could power inexpensive network terminals, PDAs, and other information appliances, to take advantage of transportable Java applications.
Many people see Java as part of a trend toward cheap, Net-based, "operating system-less" appliances that will extend the Net into more and more consumer-related areas. Only time will tell what people will do with Java, but it's probably worth at least a passing thought that the applet you write today might well be running on someone's wristwatch tomorrow. If that seems too futuristic, remember that you can already get a "smart card" (essentially a credit card) that has a Java interpreter embedded in it. Such a card could do everything from financial transactions (paying a hotel bill) to unlocking a door (the door to your hotel room) to rerouting phone calls (so your hotel room receives your business calls). The card is already here; it won't be long before the rest of the software has been built. A Java wristwatch is certainly not far away.