---

What is Portability?

There are many different definitions for portability in the context of code design. One definition of portability is limiting yourself to functions specified in a commonly accepted standard such as the Single UNIX Specification (SUS) or the Portable Operating System Interface (POSIX). This is useful for non-GUI software, particularly when moving among UNIX-based and UNIX-like systems, but it does not address graphical user interfaces and does not allow you to take advantage of operating-system–specific capabilities.

The alternative to such limitations is to design your code in a modular fashion so that additional features specific to a given OS can be “plugged into” your application. Whether you do this with a plug-in interface, a class hierarchy, or simple abstraction, this is the most effective way to allow your software to be easily ported to multiple platforms without sacrificing functionality.

There are many different ways to achieve this degree of portability, all of which have valid uses. The goal of this chapter is to describe some of these and to explain when it is appropriate (or inappropriate) to use each one. This should help you avoid some of the common pitfalls when modifying software to support multiple platforms.

Of course, the proper time to make such decisions is before you begin writing the first line of code. Since this is not always possible, this chapter will also describe ways to retrofit these concepts into an existing application in the most maintainable manner possible.

---