| English is a powerful language. With it, we can write plays and sonnets, grocery lists and business plans, love notes and contracts. A self-evident example is this book, which uses the English language to explain how to write parsers using the Java computer language. If computers could understand English, we would have little need for Java or any other programming language. Perhaps there is some potential to invent a human language that is less flexible and less ambiguous than English ”for critical tasks such as launching a spacecraft ”but people generally thrive using natural languages, applying the flexibility and even the ambiguity of language to good purpose. With computers, however, this does not work because they require precise commands to execute correctly. Computers understand very little, and arguably nothing at all. They can add numbers and move strings of text but cannot in themselves understand the idea of, say, doing something 10 times. So on the one hand we have English, which is enormously expressive, and on the other hand we have computers, which understand almost nothing. This is why programming languages emerge. A programming language such as Java is a compromise between the expressive eloquence of English and the primitive receptive abilities of a computer. Most of the words in Java are English words, and these words typically retain their English meaning. For example, while in Java has essentially the same meaning as the word while in an English sentence . Other words ( public , for example) have meaning that is specific to a programming concept but is still related to the same word in English. Java is as English-like as it can be, but Java focuses on its task of enabling us to command computers. It is reasonable to ask what the ideal programming language is, and indeed this question leads to progress and new languages such as Java. You might agree that Java is a better language because it is object-oriented. People tend to mentally model the world in terms of objects, and Java's object orientation makes it a ready receptacle for these thoughts. It simplifies the connections between, say, a real furnace and a Furnace object, and that in turn simplifies the control of a real furnace from a Java program. Object-oriented languages ease the programmer's task of interacting with objects in the real world. In addition to its object orientation, Java has many other features that make it the right choice of a computer language for many applications. Java is an excellent choice of language when the problem at hand requires giving computers precise commands. However, if you want to describe what a Web page should look like, HTML (Hypertext Markup Language) and XML (Extensible Markup Language) are more appropriate choices. XML is different from Java ”a different compromise between human understanding and computer understanding. The ideal compromise between English and computer languages depends on the context and purpose of what you are doing with the computer. This is why every programmer can benefit from learning how to create new languages. The point of learning to write parsers is that you can craft a language that fits the context you are working in. You can write a search language that is easier to use than SQL (Structured Query Language) and more specific to the data in your database. You can write a language that lets you command a robot, or one that lets you move an order through its workflow. You can create a privilege language for logically modeling which users should have access to which transactions in a system. Whatever your application is, parsers let you concoct an English-like interface to it. Parsers help computers, which work with objects, to cooperate with people, who read and write text. In practice, particularly in Java-based parsers, this implies that parsers translate text into objects. For example, a parser can translate a textual command for a robot into a command object. Another parser might translate a textual description of a product into an object that represents the product. A query parser translates textual queries into commands that a query engine understands. Parsers, then, translate between text and objects, letting you trade text for objects and objects for text, in a way that exactly suits the domain you are working in. As a language designer, you will craft a compromise between a language that is easy for humans to use and a language that computers can understand. It is well worth learning to write parsers because you will have many chances in your career to orchestrate how humans and computers interact using text. |
