Alternative Technologies

As its title suggests, this tutorial focuses on CGI programs written in Perl. Because Perl and CGI are so often used together, some people are unclear about the distinction. Perl is a developing language, and CGI is an interface that a program uses to handle requests from a web server. There are alternatives both to CGI and to Perl: there are new alternatives to CGI for handling dynamic requests, and CGI applications can be written in a variety of languages.

Why Perl?

Although CGI applications can be written in any almost any language, Perl and CGI scripting have become synonymous to many developers. As Hassan Schroeder, Sun's first webmaster, said in his oft-quoted statement, "Perl is the duct tape of the Internet." Perl is by far the most widely used language for CGI developing, and for many good reasons:

• Perl is easy to learn because it resembles other popular languages (such as C), because it is forgiving, and because when an error occurs it provides specific and detailed error messages to help you locate the problem quickly.
• Perl allows rapid development because it is interpreted; the source code does not need to be compiled before execution.
• Perl is easily portable and available on many platforms.
• Perl contains extremely powerful string manipulation operators, with regular expression matching and substitution built right into the language.
• Perl handles and manipulates binary data just as easily as it handles text.
• Perl does not require strict variable types; numbers, strings, and booleans are simply scalars.
• Perl interfaces with external applications very easily and provides its own filesystem functions.
• There are countless open source modules for Perl available on CPAN, ranging from modules for creating dynamic graphics to interfacing with Internet servers and database engines. For more information on CPAN, refer to Appendix B, "Perl Modules".

Furthermore, Perl is fast. Perl isn't strictly an interpreted language. When Perl reads a source file, it actually compiles the source into low-level opcodes and then executes them. You do not generally see compilation and execution in Perl as separate steps because they typically occur together: Perl launches, reads a source file, compiles it, runs it, and exits. This process is repeated each time a Perl script is executed, including each time a CGI script is executed. Because Perl is so efficient, however, this process occurs fast enough to handle requests for all but the most heavily trafficked websites. Note that this is considerably less efficient on Windows systems than on Unix systems because of the additional overhead that creating a new process on Windows entails.

Alternatives to CGI

Several alternatives to CGI have appeared in recent years. They all build upon CGI's legacy and provide their own approaches to the same underlying goal: responding to queries and presenting dynamic content via HTTP. Most of them also attempt to avoid the main drawback to CGI scripts: creating a separate process to execute the script every time it is requested. Others also try to make less of a distinction between HTML pages and code by moving code into HTML pages. We'll discuss the theories behind this approach in "HTML Templates". Here is a list of some of the major alternatives to CGI:

• ASP
• Active Server Pages, or ASP, was created by Microsoft for its web server, but it is now available for many servers. The ASP engine is integrated into the web server so it does not require an additional process. It allows developers to mix code within HTML pages instead of writing separate programs. As we'll see in "HTML Templates", there are modules available that allow us to do similar things using CGI. ASP supports multiple languages; the most popular is Visual Basic, but JavaScript is also supported, and ActiveState offers a version of Perl that can be used on Windows with ASP. There is also a Perl module, Apache::ASP, that supports ASP with mod_perl.
• PHP
• PHP is a developing language that is similar to Perl, and its interpreter is embedded within the web server. PHP supports embedded code within HTML pages. PHP is supported by the Apache web server.
• ColdFusion
• Allaire's ColdFusion creates more of a distinction than PHP between code pages and HTML pages. HTML pages can include additional tags that call ColdFusion functions. A number of standard functions are available with ColdFusion, and developers can create their own controls as extensions. ColdFusion was originally written for Windows, but versions for various Unix platforms are now available as well. The ColdFusion interpreter is integrated into the web server.
• Java servlets
• Java servlets were created by Sun. Servlets are similar to CGI scripts in that they are code that creates documents. However, servlets, because they use Java, must be compiled as classes before they are run, and servlets are dynamically loaded as classes by the web server when they are run. The interface is quite different than CGI. JavaServer Pages, or JSP, is another technology that allows developers to embed Java in web pages, much like ASP.
• FastCGI
• FastCGI maintains one or more instances of perl that it runs continuously along with an interface that allows dynamic requests to be passed from the web server to these instances. It avoids the biggest drawback to CGI, which is creating a new process for each request, while still remaining largely compatible with CGI. FastCGI is available for a variety of web servers. We'll discuss FastCGI further in "Efficiency and Optimization".
• mod_perl
• mod_perl is a module for the Apache web server that also avoids creating separate instances of perl for each CGI. Instead of maintaining a separate instance of perl like FastCGI, mod_perl embeds the perl interpreter inside the web server. This gives it a performance advantage and also gives Perl code written for mod_perl access to Apache's internals. We'll discuss mod_perl further in "Efficiency and Optimization".

Despite a proliferation of these competing technologies, CGI continues to be the most popular method for delivering dynamic pages, and, despite what the marketing literature for some of its competitors may claim, CGI will not go away any time soon. Even if you do imagine that you may begin using other technologies down the road, learning CGI is a valuable investment. Because CGI is such a thin interface, learning CGI teaches you how web transactions works at a basic level, which can only further your understanding of other technologies built upon this same foundation. Additionally, CGI is universal. Many alternative technologies require that you install a particular combination of technologies in addition to your web server in order to use them. CGI is supported by virtually every web server "right out of the box" and will continue to be that way far into the future.