If you are familiar with other language compilers—such as Microsoft Visual
C++, Borland Turbo Pascal, or any of the various assembly language compilers available
for a given processor chip—you know that a language compiler turns a programmer’s
source code into native machine code that is linked into an executable program
file.
Before version 5 of VB, the Visual Basic compiler did not generate the machine
code common to other compilers. Instead, it could only create something called
pseudocode, commonly abbreviated to P-Code.
To understand how P-Code differs from machine code, let’s take a look
at the difference between programs based on interpreted languages and programs
based on compiled languages.
At some level, every computer program can be said to consist of nothing but
the source code used to write it. Programmers write statements in a high-level
language (such as Visual Basic, COBOL, or C++). If you assemble a series of these
statements that collectively are supposed to do something useful, you have a computer
program. Before the computer can do anything with a program, however, these high-level
language statements must be translated into something the computer can understand:
machine code.
Machine code instructions govern the most fundamental tasks performed by a
CPU. For Intel chips, those basic operations include things such as data transfer,
arithmetic, bit manipulation, string manipulation, control transfer, and flag
control. The format of an instruction falls into two parts; One part identifies
the operation code, while the other identifies the address in the computer’s
memory at which the data to be used in the operation is stored.
Machine code varies with each processor, however. The Motorola chips used in
Apple computers, for example, don’t respond to the same set of instructions
used by the Intel chips found in computers that run Microsoft Windows. For a given
program to run on a computer, it must be converted into the machine code appropriate
for that computer. The code used by a particular processor chip is also known
as its native machine code.
Until a program is translated from the language used by the programmer into
native code, the computer can’t do anything with the program. The difference
between an interpreted program and a compiled program is the point at which this
translation occurs.
With a compiled program, the process of producing an executable file from your
source code takes two basic steps. The first step is the compile step. If you
choose to compile to native code, the compiler produces a series of intermediate
files from your source code. These intermediate files are commonly called object
files, and many compilers (including the VB compiler) give these intermediate
files an extension of OBJ.
Even though they consist of machine code, the object files themselves can’t
be used directly by your computer. Because your computer relies on an operating
system (for example, Microsoft Windows 9x or Windows NT), another step—called
the link step—is necessary to produce an EXE file. During the link step,
the object files produced by the compiler are linked together with some startup
code that tells your operating system where to load your program into memory.
The result is written to disk in a form that your computer can use.
