File: gawk.info,  Node: Basic High Level,  Next: Basic Data Typing,  Up: Basic Concepts

D.1 What a Program Does
=======================

At the most basic level, the job of a program is to process some input
data and produce results.  See *note Figure D.1: figure-general-flow.


                  _______
+------+         /       \         +---------+
| Data | -----> < Program > -----> | Results |
+------+         \_______/         +---------+

Figure D.1: General Program Flow

   The "program" in the figure can be either a compiled program(1) (such
as 'ls'), or it may be "interpreted".  In the latter case, a
machine-executable program such as 'awk' reads your program, and then
uses the instructions in your program to process the data.

   When you write a program, it usually consists of the following, very
basic set of steps, as shown in *note Figure D.2: figure-process-flow.:


                              ______
+----------------+           / More \  No       +----------+
| Initialization | -------> <  Data  > -------> | Clean Up |
+----------------+    ^      \   ?  /           +----------+
                      |       +--+-+
                      |          | Yes
                      |          |
                      |          V
                      |     +---------+
                      +-----+ Process |
                            +---------+

Figure D.2: Basic Program Steps

Initialization
     These are the things you do before actually starting to process
     data, such as checking arguments, initializing any data you need to
     work with, and so on.  This step corresponds to 'awk''s 'BEGIN'
     rule (*note BEGIN/END::).

     If you were baking a cake, this might consist of laying out all the
     mixing bowls and the baking pan, and making sure you have all the
     ingredients that you need.

Processing
     This is where the actual work is done.  Your program reads data,
     one logical chunk at a time, and processes it as appropriate.

     In most programming languages, you have to manually manage the
     reading of data, checking to see if there is more each time you
     read a chunk.  'awk''s pattern-action paradigm (*note Getting
     Started::) handles the mechanics of this for you.

     In baking a cake, the processing corresponds to the actual labor:
     breaking eggs, mixing the flour, water, and other ingredients, and
     then putting the cake into the oven.

Clean Up
     Once you've processed all the data, you may have things you need to
     do before exiting.  This step corresponds to 'awk''s 'END' rule
     (*note BEGIN/END::).

     After the cake comes out of the oven, you still have to wrap it in
     plastic wrap to keep anyone from tasting it, as well as wash the
     mixing bowls and utensils.

   An "algorithm" is a detailed set of instructions necessary to
accomplish a task, or process data.  It is much the same as a recipe for
baking a cake.  Programs implement algorithms.  Often, it is up to you
to design the algorithm and implement it, simultaneously.

   The "logical chunks" we talked about previously are called "records",
similar to the records a company keeps on employees, a school keeps for
students, or a doctor keeps for patients.  Each record has many
component parts, such as first and last names, date of birth, address,
and so on.  The component parts are referred to as the "fields" of the
record.

   The act of reading data is termed "input", and that of generating
results, not too surprisingly, is termed "output".  They are often
referred to together as "input/output," and even more often, as "I/O"
for short.  (You will also see "input" and "output" used as verbs.)

   'awk' manages the reading of data for you, as well as the breaking it
up into records and fields.  Your program's job is to tell 'awk' what to
do with the data.  You do this by describing "patterns" in the data to
look for, and "actions" to execute when those patterns are seen.  This
"data-driven" nature of 'awk' programs usually makes them both easier to
write and easier to read.

   ---------- Footnotes ----------

   (1) Compiled programs are typically written in lower-level languages
such as C, C++, or Ada, and then translated, or "compiled", into a form
that the computer can execute directly.