Among the legal remedies being considered in the Microsoft antitrust trial is a court order that would force the company to reveal the “source code” for part or all of its Windows operating system. What exactly does this mean?
Source code is simply the human-readable version of the computer commands that make up a program. It is the letters and symbols that software engineers type into their computers when they create an application or operating system. For example, if a C++ programmer wrote a program to make his computer display the words, “Hello, World,” the source code might look like this:
#include
void main()
{
cout<<”Hello, World”;
}
Although these commands are intelligible to engineers, they are useless to computers, which understand only ones and zeros. So, to make the source code into a functioning program, translation software (called a “compiler”) must convert it into the binary “object code” that computers can process.
Open source software, such as the Linux operating system, makes both the object code and the source code freely available so that any programmer can figure out how the program is constructed and then alter it to his satisfaction. But closed source software, such as Windows and most commercial programs, is shipped with the object code only. Software companies argue that source code contains valuable intellectual property that must be protected. One common analogy compares source code to the formula for Coca-Cola: The Coca-Cola Co. ships only bottles of Coke to its customers, not the formula for the drink, because it doesn’t want competitors to brew knock-offs or use the formula to improve their drinks.
With software, “decompilers”–or, reverse-translation programs–can theoretically convert object code back into source code. But their inaccuracies render them largely ineffective. (Think of one person translating a sentence from English to Japanese and another translating it back into English; chances are slim that the sentence would end up in its original form.) Additionally, attempting to reverse engineer software is illegal: Like most commercial software companies, Microsoft forbids it as part of its licensing agreement.
If Microsoft (which owns Slate) were to make the Windows source code public, competitors could use this information to market Windows clones or alter their operating systems (Linux, Mac OS, etc.) to run software designed for Windows. However, neither project would be easy. Windows 98 contains approximately 20 million lines of source code written in multiple computer languages. To use the source code effectively, a competitor would need to figure out which commands affect which functions and how changes to one part would affect the functioning of the whole–a nearly impossible guessing game for even the most experienced programmers. For this reason, releasing the Windows source code might not instantly erode Microsoft’s dominance in PC operating systems unless a great number of software engineers applied themselves to the effort. Microsoft argues that releasing the source code would be bad for consumers because it would result in many incompatible versions of Windows.
So far, the results of releasing source code to the public are mixed. It can stir great creativity among programmers, as the Linux story illustrates. Or, it can prove a disappointment. Netscape made public the source code for its Navigator browser a year an a half ago, and most estimate that it will take at least another six months for programmers to create a new product with it. And a browser’s source code is only a fraction of the size of that of an operating system.
