Stylish design

1 Comment Written by Robert on April 15, 2008 – 11:37 am

Its use far transcends its deployment as a convenient language for CGI scripting. Perl is primarily an ‘extraction and reporting’ language. As such, it should be part of your set of programming tools. While you may want to focus on web applications, you are still going to need to analyze large data sets, such as the log files from your Apache server, and produce reports that abstract the important data. Another reason to learn Perl is to get to use its ‘regular expression matcher’; if you haven’t previously encountered regular expressions, now is the time to learn.

The basic control structures will be familiar to all who have programmed in C or similar languages. Perl has some somewhat unusual data types – lists and hashes. In some respects these are a little like classes: you create lists and hashes; these instances hold data; and there are defined functions that operate on these structures. However, the syntax for code using these basic lists and hashes does not utilize class/object/method conventions. (Perl does have extensions that support real object-oriented- style programming, but these are beyond the scope of this introductory treatment.)

One of Perl’s great strengths is its ability to extract sub-strings from text, and to search text to find sub-strings with particular patterns; an example use might be finding links in an HTML document (links are sub-strings that have the general pattern <a href=something other-stuff> some text </a>). Perl’s ‘regular expression’ matching functions and operators provide a convenient basis for the construction of such data extraction programs.

Perl has good interfaces to the underlying operating system and file system. Perl has many extension libraries. DBI provides elements that will be familiar to those who have worked with Java’s JDBC; the names may have changed, but it is again essentially a matter of using Connection, Statement and ResultSet objects.

Simple CGI programs can be written using standard Perl, but there are extension libraries that supply helper objects and functions that may be useful in more demanding applications.

Perl’s ancestry stems from the tools conventionally employed on Unix systems. Some of these Unix tools data back to the mid-1970s, but still remain in use. The following are among the more important:

ed (vi) text editor

The ed editor (and its descendants, such as vi) incorporates a regular expression matcher that is used for sophisticated global search and replacement operations.

sed

sed is a ‘script editor’. It can read a control file with ed-style matching and replacement commands, and then apply these commands to input text data file(s).

awk

awk’s role is ‘pattern scanning and replacement’. An awk program reads text files line-byline. Each line is compared with one or more awk patterns; if a line matches a pattern, a particular action is performed. The action might extract data, or simply replace matched text and output the modified line.

sh

The Unix shell (command interpreter) is itself a powerful tool. sh (or descendants such as bash, ksh csh) allows for looping constructs and command invocations, and of course it supports numerous operations on files and directories. The shell is an infinitely extensible programming language: shell scripts can be saved to files and these files can be marked as ‘executable’; these new executables become commands that extend the language.

Generations of Unix programmers have exploited these tools to build complex data processing systems. Shell scripts formed the basic programming skeleton; sophisticated data processing operations were handled through sed and awk scripts. Such script programs are somewhat more limited than C programming, but generally it is quicker to implement a script than to write a new C program for a given task.

Larry Wall, who developed Perl, had grown up with these Unix tools and C programming. He liked the power of C (its high ‘manipulexity’ rating), and the speed of scripting with sed and awk (their high ‘whipitupitude’ ratings). What he desired was something that would combine high manipulexity with high whipitupitude; something that, using more conventional programming idioms, could accomplish tasks commonly handled through sed or awk scripts.

Wall composed his new tool from Unix ingredients. From C, he took control structures, I/O libraries and system calls, but left out ‘functionitis’ and limitations on strings. From sed, he took substitution and translation operations, but left out ‘impenetrability’ and limitations. awk gave him associations and strings; he went without awk’s sloth, and control limitations. From sh he took processes, lists and interpolations, but forsook features like backslashitis, list-string confusions and other limitations. This brew yielded Perl – a Practical Extraction and Reporting Language.

Perl is essentially an interpreted language. The interpreter reads the Perl program (script); checks the syntax, and converts the code into an internal representation that allows faster processing at run time. The converted Perl program is then run. The typical Perl program performs text processing – finding and replacing strings, counting symbols, and so forth. Perl programs will typically perform almost as well as C programs written for the same tasks. Perl’s interpreter is in C, and C’s stdio I/O libraries are used for reading and writing data. Consequently, the I/O operations of a Perl script are as efficient as those of C programs. Most Perl applications are I/O bound. Any loss in performance from Perl’s interpreted data processing code is likely to be small; most of the work will involve I/O operations.

The interpreter defines the core of the Perl language. But this core can be extended through libraries implemented in Perl. Perl is essentially ‘open source’. Many extension libraries were contributed as Perl evolved. Contributions were made to organizations like the Comprehensive Perl Archive Network or to the O’Reilly publishing company which hosts perl.com. Naturally, every contributor has had his or her own favorite programming idioms and function libraries. All have been incorporated, making the evolved Perl quite an eclectic mix. One of the favorite slogans of the Perl community is ‘there is always another way’.

Popularity: 61% [?]