Perl

2007 Schools Wikipedia Selection. Related subjects: Computer Programming

Perl

Paradigm:	Multi-paradigm
Appeared in:	1987
Designed by:	Larry Wall
Latest release:	5.8.8 / January 31, 2006
Typing discipline:	Dynamic
Influenced by:	AWK, BASIC-PLUS, C, C++, Lisp, Pascal, sed, Unix shell
Influenced:	Python, PHP, Ruby
OS:	Cross-platform
License:	GNU General Public License, Artistic License
Website:	http://www.perl.org/

Perl is a dynamic programming language designed by Larry Wall and first released in 1987. Perl borrows features from a variety of other languages including C, shell scripting ( sh), AWK, sed and Lisp.

Structurally, Perl is based on the brace-delimited block style of AWK and C, and was widely adopted for its strengths in string processing, and lack of the arbitrary limitations of many scripting languages at the time.

History

Wall began work on Perl in 1987, while working as a programmer at Unisys, and released version 1.0 to the comp.sources.misc newsgroup on December 18, 1987. The language expanded rapidly over the next few years. Perl 2, released in 1988, featured a better regular expression engine. Perl 3, released in 1989, added support for binary data.

Until 1991, the only documentation for Perl was a single (increasingly lengthy) man page. In 1991, Programming Perl (the Camel Book) was published, and became the de facto reference for the language. At the same time, the Perl version number was bumped to 4, not to mark a major change in the language, but to identify the version that was documented by the book.

Perl 4 went through a series of maintenance releases, culminating in Perl 4.036 in 1993. At that point, Larry Wall abandoned Perl 4 to begin work on Perl 5. Perl 4 remains at version 4.036 to this day.

Development of Perl 5 continued into 1994. The perl5-porters mailing list was established in May 1994 to coordinate work on porting Perl 5 to different platforms. It remains the primary forum for development, maintenance, and porting of Perl 5.

Perl 5 was released on October 17, 1994. It was a nearly complete rewrite of the interpreter, and added many new features to the language, including objects, references, packages, and modules. Importantly, modules provided a mechanism for extending the language without modifying the interpreter. This allowed the core interpreter to stabilize, even as it enabled ordinary Perl programmers to add new language features.

On October 26, 1995, the Comprehensive Perl Archive Network (CPAN) was established. The CPAN is a collection of web sites that archive and distribute Perl sources, binary distributions, documentation, scripts, and modules.

As of 2006, Perl 5 is still being actively maintained. Important features and some essential new language constructs have been added along the way, including Unicode support, threads, an improved support for object oriented programming and many other enhancements. The latest stable release is Perl 5.8.8.

Name

Perl was originally named "Pearl", after the Parable of the Pearl. Larry Wall wanted to give the language a short name with positive connotations; he claims that he looked at (and rejected) every three- and four-letter word in the dictionary. He also considered naming it after his wife Gloria. Wall discovered the existing PEARL programming language before Perl's official release and changed the spelling of the name.

The name is normally capitalized (Perl) when referring to the language and uncapitalized (perl) when referring to the interpreter program itself since Unix-like file systems are case sensitive. Before the release of the first edition of Programming Perl it was common to refer to the language as perl; Randal L. Schwartz, however, capitalised the language's name in the book to make it stand out better when typeset. The case distinction was subsequently adopted by the community.

The name is occasionally given as "PERL" (for Practical Extraction and Report Language). Although the expansion has prevailed in many of today's manuals, including the official Perl man page, it is a backronym and officially the name stands for nothing. The spelling of PERL in all caps is therefore used as a shibboleth for detecting community outsiders. Several other backronyms have been suggested, including the humorous Pathologically Eclectic Rubbish Lister.

The camel symbol

Perl is generally symbolized by a camel, which was a result of the picture chosen by publisher O'Reilly Media for the cover of Programming Perl, which consequently acquired the name The Camel Book. O'Reilly owns the symbol as a trademark, but claims to use their legal rights only to protect the "integrity and impact of that symbol" . O'Reilly allows non-commercial use of the symbol, and provides Programming Republic of Perl logos (see above) and Powered by Perl buttons.

Overview

Perl is a general-purpose programming language originally developed for text manipulation and now used for a wide range of tasks including system administration, web development, network programming, GUI development, and more.

The language is intended to be practical (easy to use, efficient, complete) rather than beautiful (tiny, elegant, minimal). Its major features include support for multiple programming paradigms ( procedural, object-oriented, and functional styles), automatic memory management, built-in support for text processing, and a large collection of third-party modules.

Features

The overall structure of Perl derives broadly from C. Perl is procedural in nature, with variables, expressions, assignment statements, brace-delimited code blocks, control structures, and subroutines.

Perl also takes features from shell programming. All variables are marked with leading sigils, which unambiguously identify the data type (scalar, array, hash, etc.) of the variable in context. Importantly, sigils allow variables to be interpolated directly into strings. Like the Unix shells, Perl has many built-in functions for common tasks, like sorting, and for accessing system facilities.

Perl takes lists from Lisp, associative arrays (hashes) from AWK, and regular expressions from sed. These simplify and facilitate many parsing, text handling, and data management tasks.

In Perl 5, features were added that support complex data structures, first-class functions (i.e. closures as values), and an object-oriented programming model. These include references, packages, class-based method dispatch, and lexically scoped variables, along with compiler directives (for example, the strict pragma). A major additional feature introduced with Perl 5 was the ability to package code as reusable modules. Larry Wall later stated that "The whole intent of Perl 5's module system was to encourage the growth of Perl culture rather than the Perl core."

All versions of Perl do automatic data typing and memory management. The interpreter knows the type and storage requirements of every data object in the program; it allocates and frees storage for them as necessary. Legal type conversions are done automatically at run time; illegal type conversions are fatal errors.

Design

The design of Perl can be understood as a response to three broad trends in the computer industry: falling hardware costs, rising labor costs, and improvements in compiler technology. Many earlier computer languages, such as Fortran and C, were designed to make efficient use of expensive computer hardware. In contrast, Perl is designed to make efficient use of expensive computer programmers.

Perl has many features that ease the programmer's task at the expense of greater CPU and memory requirements. These include automatic memory management; dynamic typing; strings, lists, and hashes; regular expressions; introspection and an eval() function.

Wall was trained as a linguist, and the design of Perl is very much informed by linguistic principles. Examples include Huffman coding (common constructions should be short), good end-weighting (the important information should come first), and a large collection of language primitives. Perl favors language constructs that are natural for humans to read and write, even where they complicate the Perl interpreter.

Perl syntax reflects the idea that "things that are different should look different". For example, scalars, arrays, and hashes have different leading sigils. Array indices and hash keys use different kinds of braces. Strings and regular expressions have different standard delimiters. This approach can be contrasted with languages like Lisp, where the same S-expression construct and basic syntax is used for many different purposes.

Perl does not enforce any particular programming paradigm (procedural, object-oriented, functional, etc.), or even require the programmer to choose among them.

There is a broad practical bent to both the Perl language and the community and culture that surround it. The preface to Programming Perl begins, "Perl is a language for getting your job done." One consequence of this is that Perl is not a tidy language. It includes features if people use them, tolerates exceptions to its rules, and employs heuristics to resolve syntactical ambiguities. Because of the forgiving nature of the compiler, bugs can be hard to find sometimes. Discussing the variant behaviour of built-in functions in list and scalar contexts, the perlfunc(1) manual page says "In general, they do what you want, unless you want consistency."

Perl has several mottos that convey aspects of its design and use. One is " There's more than one way to do it." (TMTOWTDI, usually pronounced 'Tim Toady'). Others are "Perl: the Swiss Army Chainsaw of Programming Languages" and "No unnecessary limits". A stated design goal of Perl is to make easy tasks easy and difficult tasks possible. Perl has also been called "The Duct Tape of the Internet".

Applications

Perl has many and varied applications, compounded by the availability of many standard and third-party modules.

Perl has been used since the early days of the Web to write CGI scripts. It is known as one of "the three Ps" (Perl, Python and PHP), which are the most popular scripting languages for generating Web applications, and is an integral component of the popular LAMP solution stack for web development. Large projects written in Perl include Slash, IMDb and UseModWiki, an early, influential wiki engine. Many high-traffic websites, such as Amazon.com and Ticketmaster.com use Perl extensively.

Perl is often used as a glue language, tying together systems and interfaces that were not specifically designed to interoperate, and for "data munging", converting or processing large amounts of data for tasks like creating reports. In fact, these strengths are intimately linked. The combination makes perl a popular all-purpose tool for system administrators, particularly as short programs can be entered and run on a single command line.

Perl is also widely used in finance and bioinformatics, where it is valued for rapid application development and deployment, and the ability to handle large data sets.

Implementation

Perl is implemented as a core interpreter, written in C, together with a large collection of modules, written in Perl and C. The source distribution is, as of 2005, 12 MB when packaged in a tar file and compressed. The interpreter is 150,000 lines of C code and compiles to a 1 MB executable on typical machine architectures. Alternatively, the interpreter can be compiled to a link library and embedded in other programs. There are nearly 500 modules in the distribution, comprising 200,000 lines of Perl and an additional 350,000 lines of C code. Much of the C code in the modules consists of character encoding tables.

The interpreter has an object-oriented architecture. All of the elements of the Perl language—scalars, arrays, hashes, coderefs, file handles—are represented in the interpreter by C structs. Operations on these structs are defined by a large collection of macros, typedefs and functions; these constitute the Perl C API. The Perl API can be bewildering to the uninitiated, but its entry points follow a consistent naming scheme, which provides guidance to those who use it.

The execution of a Perl program divides broadly into two phases: compile-time and run-time. At compile time, the interpreter parses the program text into a syntax tree. At run time, it executes the program by walking the tree. The text is parsed only once, and the syntax tree is subject to optimization before it is executed, so the execution phase is relatively efficient. Compile-time optimizations on the syntax tree include constant folding and context propagation, but peephole optimization is also performed. However, compile-time and run-time phases may nest: BEGIN code blocks execute at compile-time, while the eval function initiates compilation during runtime. Both operations are implicit in a number of others - most notably, the use clause that loads libraries, known in Perl as modules, implies a BEGIN block.

Perl is a dynamic language and has a context-sensitive grammar which can be affected by code executed during an intermittent run-time phase. (See examples. ) Therefore Perl cannot be parsed by a straight Lex/ Yacc lexer/parser combination. Instead, the interpreter implements its own lexer, which coordinates with a modified GNU bison parser to resolve ambiguities in the language. It is said that "only perl can parse Perl", meaning that only the Perl interpreter (perl) can parse the Perl language (Perl). The truth of this is attested to by the persistent imperfections of other programs that undertake to parse Perl, such as source code analyzers and auto-indenters, which have to contend not only with the many ways to express unambiguous syntactic constructs, but also the fact that Perl cannot be parsed in the general case without executing it.

Maintenance of the Perl interpreter has become increasingly difficult over the years. The code base has been in continuous development since 1994. The code has been optimized for performance at the expense of simplicity, clarity, and strong internal interfaces. New features have been added, yet virtually complete backward compatibility with earlier versions is maintained. The size and complexity of the interpreter is a barrier to developers who wish to work on it.

Perl is distributed with some 120,000 functional tests. These run as part of the normal build process, and extensively exercise the interpreter and its core modules. Perl developers rely on the functional tests to ensure that changes to the interpreter do not introduce bugs; conversely, Perl users who see the interpreter pass its functional tests on their system can have a high degree of confidence that it is working properly.

There is no written specification or standard for the Perl language, and no plans to create one for the current version of Perl. There has only ever been one implementation of the interpreter. That interpreter, together with its functional tests, stands as a de facto specification of the language.

Availability

Perl is free software, and is licensed under both the Artistic License and the GNU General Public License. Distributions are available for most operating systems. It is particularly prevalent on Unix and Unix-like systems, but it has been ported to most modern (and many obsolete) platforms. With only six reported exceptions, Perl can be compiled from source code on all Unix-like, POSIX-compliant or otherwise Unix-compatible platforms. However, this is rarely necessary, as Perl is included in the default installation of many popular operating systems.

Because of special changes required to support Mac OS Classic, a special port called MacPerl was shipped independently.

Windows

Users of Microsoft Windows typically install a native binary distribution of Perl. Compiling Perl from source code under Windows is possible, but most installations lack the requisite C compiler.

The Cygwin emulation layer provides another way of running Perl under Windows. Cygwin provides a Unix-like environment on Windows that includes gcc, so compiling Perl from source is a more accessible option for users who take this approach.

In June 2006, win32.perl.org was launched by Adam Kennedy on behalf of the The Perl Foundation. It is a community website for "all things Windows and Perl."

Language structure

In Perl, the canonical Hello world program is ocassionally stated as:

#!/usr/bin/perl -w
use strict;
print "Hello, world!\n";    # "\n" is a 'newline'

The first line is the shebang, which tells the operating system where to find the Perl interpreter. The second line introduces the strict pragma which is used in many large software projects for quality control. The third prints the string Hello, world! and a newline. A comment ( '\n' is a 'newline' ) follows.

The # sign on the third line is a 'comment token', which allows the perl interpreter to ignore everything after the # sign, up to the end of the line of code.

The shebang is the usual way to invoke the interpreter on Unix systems. Windows systems may rely on the shebang, or they may associate a .pl file extension with the Perl interpreter. Some text editors also use the shebang line as a hint about what mode to operate in. If the program is executed by perl and not invoked via the shell, the line starting with the shebang is parsed for options, and otherwise ignored. For details see the perlrun manpage.

It should be noted that, as the Perl "Hello world" program requires no variables, subroutines or anything else that would even potentially violate the strict pragma without causing a fatal error without using strict, and as the use of strict restrictions is not a requirement for a "Hello world" program, the only reason to use strict in it is to avoid being potentially yelled at by other programmers for not using strict.

Further, as it's perfectly possible to invoke the perl interpreter directly from a command line, and additionally as the 'shebang' line is not a part of Perl but a requirement of some shells (and treated as a comment by perl), the shebang line is also not actually necessary. Indeed, in most Win32 implementations it's completely useless.

Finally, there are no need for explanatory comments inside a demonstrative "Hello world" program. Additionally, the final line of any block of code in a Perl program, which includes the implied block around the entire script, does not actually need a semicolon terminator.

Thus, in reality, in Perl, the canonical Hello world program is

print "Hello, world!\n"

and that's all.

Data types

Perl has four fundamental data types: scalars, lists, hashes and filehandles:

• A scalar is a single value; it may be a number, a string or a reference
• A list is an ordered collection of scalars (a variable that holds a list is called an array)
• A hash, or associative array, is a map from strings to scalars; the strings are called keys and the scalars are called values.
• A filehandle is a map to a file, device, or pipe which is open for reading, writing, or both.

All variables are marked by a leading sigil, which identifies the data type being accessed (not the type of the variable itself), except filehandles, which aren't. The same name may be used for variables of different types, without conflict.

 $foo   # a  scalar
 @foo   # an  array
 %foo   # a  hash
 foo    # a  Filehandle, but nice programmers use FOO, not foo.

Numbers are written in the usual way; strings are enclosed by quotes of various kinds.

 $n      = 42;
 $name   = "joe";
 $colour  = 'red';
 $animal = qq!frog!;

Perl will convert strings into numbers and vice versa depending on the context in which they are used. In the following example the strings $n and $m are treated as numbers when they are the arguments to the addition operator. This code prints the number '5', discarding non number information for the operation, although the variable values remain the same. (The string concatenation operator is not +, but .)

 $n     = "3 apples";
 $m     = "2 oranges";
 print $n + $m;

Perl also has a boolean context that it uses in evaluating conditional statements. The following values all evaluate as false in Perl:

 $false = 0;     # the number zero
 $false = 0.0;   # the number zero as a float
 $false = '0';   # the string zero
 $false = "";    # the empty string
 $false = undef; # the return value from undef

All other values are evaluated to true. This includes the odd self-describing string of "0 but true", which in fact is 0 as a number, but true when used as a boolean. (Any non-numeric string would also have this property, but this particular string is ignored by Perl with respect to numeric warnings.) A less explicit but more conceptually portable version of this string is '0E0' or '0e0', which does not rely on characters being evaluated as 0, as '0E0' is literally "zero to the exponent of zero."

Evaluated boolean expressions also return scalar values. Although the documentation does not promise which particular true or false is returned (and thus cannot be relied on), many boolean operators return 1 for true and the empty-string for false (which evaluates to zero in a numeric context). The defined() function tells if the variable has any value set. In the above examples defined($false) is true for every value except undef.

If a specifically 1 or 0 result (as in C) is needed, an explicit conversion is thought by some authors to be required:

 my $real_result = $boolean_result ? 1 : 0;

However, an implicit conversion can be used instead:

 my $real_result = $boolean_result + 0;

A list is written by listing its elements, separated by commas, and enclosed by parentheses where required by operator precedence.

 @scores = (32, 45, 16, 5);

Or, then again, it can be written some other half dozen ways, at least:

 @scores = qw(32 45 16 5);
 @scores = split /-/, '32-45-16-5';
 push @scores, $_ for 32, 45, 16, 5;

A hash may be initialized from a list of key/value pairs.

 %favorite = (joe => 'red',
              sam => 'blue');

Or it may simply be defined piece by piece:

 $favourite{joe} = 'red';
 $favourite{sam} = 'blue';

Individual elements of a list are accessed by providing a numerical index, in square brackets. Individual values in a hash are accessed by providing the corresponding key, in curly braces. The $ sigil identifies the accessed element as a scalar.

 $scores[2]      # an element of @scores
 $favorite{joe}  # a value in %favorite

Multiple elements may be accessed by using the @ sigil instead (identifying the result as a list).

 @scores[2, 3, 1]    # three elements of @scores
 @favorite{'joe', 'sam'} # two values in %favorite

The number of elements in an array can be obtained by evaluating the array in scalar context or with the help of the $# sigil. The latter gives the index of the last element in the array, not the number of elements.

 $count = @friends;
 $#friends       # the index of the last element in @friends
 $#friends+1     # usually the number of elements in @friends
                 # this is one more than $#friends because the first element is at
                 # index 0, not 1. Unless the programmer reset this to a
                 # different value, which most Perl manuals encourage her
                 # not to do.

There are a few functions that operate on entire hashes.

 @names     = keys   %address;
 @addresses = values %address;
 1 while ($name, $address) = each %address;

Control structures

Perl has several kinds of control structures.

It has block-oriented control structures, similar to those in the C and Java programming languages. Conditions are surrounded by parentheses, and controlled blocks are surrounded by braces:

label while ( cond ) { ... }
label while ( cond ) { ... } continue { ... }
label for ( init-expr ; cond-expr ; incr-expr ) { ... }
label foreach var ( list ) { ... }
label foreach var ( list ) { ... } continue { ... }
if ( cond ) { ... }
if ( cond ) { ... } else { ... } 
if ( cond ) { ... } elsif ( cond ) { ... } else { ... } 

Where only a single statement is being controlled, statement modifiers provide a lighter syntax:

statement if      cond ;
statement unless  cond ;
statement while   cond ;
statement until   cond ;
statement foreach list ;

Short-circuit logical operators are commonly used to effect control flow at the expression level:

expr and expr
expr or  expr

The flow control keywords next, last, return, and redo are expressions, so they can be used with short-circuit operators.

Perl also has two implicit looping constructs:

 results = grep { ... } list
 results = map  { ... } list

grep returns all elements of list for which the controlled block evaluates to true. map evaluates the controlled block for each element of list and returns a list of the resulting values. These constructs enable a simple functional programming style.

There is no switch statement (multi-way branch) in Perl 5. The Perl documentation describes a half-dozen ways to achieve the same effect by using other control structures. There is a Switch module, however, which provides functionality modeled on the forthcoming Perl 6 re-design.

Perl includes a goto label statement, but it is rarely used. Situations where a goto is called for in other languages don't occur as often in Perl due to its breadth of flow control options.

There is also a goto &sub statement that performs a tail call. It terminates the current subroutine and immediately calls the specified sub. This is used in situations where a caller can perform more efficient stack management than Perl itself (typically because no change to the current stack is required), and in deep recursion tail calling can have substantial positive impact on performance because it avoids the overhead of scope/stack management on return.

Subroutines

Subroutines are defined with the sub keyword, and invoked simply by naming them. If the subroutine in question has not yet been declared, parentheses are required for proper parsing.

foo();             # parentheses required here...
sub foo { ... }
foo;               # ... but not here

A list of arguments may be provided after the subroutine name. Arguments may be scalars, lists, or hashes.

foo $x, @y, %z;

The parameters to a subroutine need not be declared as to either number or type; in fact, they may vary from call to call. Arrays are expanded to their elements, hashes are expanded to a list of key/value pairs, and the whole lot is passed into the subroutine as one undifferentiated list of scalars.

Whatever arguments are passed are available to the subroutine in the special array @_. The elements of @_ are aliased to the actual arguments; changing an element of @_ changes the corresponding argument.

Elements of @_ may be accessed by subscripting it in the usual way.

$_[0], $_[1]

However, the resulting code can be difficult to read, and the parameters have pass-by-reference semantics, which may be undesirable.

One common idiom is to assign @_ to a list of named variables.

my($x, $y, $z) = @_;

This effects both mnemonic parameter names and pass-by-value semantics. The my keyword indicates that the following variables are lexically scoped to the containing block.

Another idiom is to shift parameters off of @_. This is especially common when the subroutine takes only one argument.

my $x = shift;

Subroutines may return values.

return 42, $x, @y, %z;

If the subroutine does not exit via a return statement, then it returns the last expression evaluated within the subroutine body. Arrays and hashes in the return value are expanded to lists of scalars, just as they are for arguments.

The returned expression is evaluated in the calling context of the subroutine; this can surprise the unwary.

sub list  {      (4, 5, 6)     }
sub array { @x = (4, 5, 6); @x }

$x = list;   # returns 6 - last element of list
$x = array;  # returns 3 - number of elements in list
@x = list;   # returns (4, 5, 6)
@x = array;  # returns (4, 5, 6)

A subroutine can discover its calling context with the wantarray function.

sub either { wantarray ? (1, 2) : "Oranges" }

$x = either;    # returns "Oranges"
@x = either;    # returns (1, 2)

Regular expressions

The Perl language includes a specialized syntax for writing regular expressions (REs), and the interpreter contains an engine for matching strings to regular expressions. The regular expression engine uses a backtracking algorithm, extending its capabilities from simple pattern matching to string capture and substitution. The regular expression engine is derived from regex written by Henry Spencer.

The Perl regular expression syntax was originally taken from Unix Version 8 regular expressions. However, it diverged before the first release of Perl, and has since grown to include many more features. Other languages and applications are now adopting Perl compatible regular expressions over POSIX regular expressions including PHP, Ruby, Java, and the Apache HTTP server.

The m// (match) operator introduces a regular expression match. (The leading m may be omitted for brevity.) In the simplest case, an expression like

 $x =~ m/abc/

evaluates to true if and only if the string $x matches the regular expression abc.

Portions of a regular expression may be enclosed in parentheses; corresponding portions of a matching string are captured. Captured strings are assigned to the sequential built-in variables $1, $2, $3, ..., and a list of captured strings is returned as the value of the match.

 $x =~ m/a(.)c/;  # capture the character between 'a' and 'c'

The s/// (substitute) operator specifies a search and replace operation:

 $x =~ s/abc/aBc/;   # upcase the b

Perl regular expressions can take modifiers. These are single-letter suffixes that modify the meaning of the expression:

 $x =~ m/abc/i;      # case-insensitive pattern match
 $x =~ s/abc/aBc/g;  # global search and replace

Regular expressions can be dense and cryptic. This is because regular expression syntax is extremely compact, generally using single characters or character pairs to represent its operations. Perl provides some relief from this problem with the /x modifier, which allows programmers to place whitespace and comments inside regular expressions:

 $x =~ m/a     # match 'a'
         .     # match any character
         c     # match 'c'
          /x;

One common use of regular expressions is to specify delimiters for the split operator:

 @words = split m/,/, $line;   # divide $line into comma-separated values

The split operator complements string capture. String capture returns the parts of a string that match a regular expression; split returns the parts that don't match.

Database interfaces

Perl is widely favored for database applications. Its text handling facilities are good for generating SQL queries; arrays, hashes and automatic memory management make it easy to collect and process the returned data.

In early versions of Perl, database interfaces were created by relinking the interpreter with a client-side database library. This was somewhat clumsy; a particular problem was that the resulting perl executable was restricted to using just the one database interface that it was linked to. Also, relinking the interpreter was sufficiently difficult that it was only done for a few of the most important and widely used databases.

In Perl 5, database interfaces are implemented by Perl DBI modules. The DBI (Database Interface) module presents a single, database-independent interface to Perl applications, while the DBD:: (Database Driver) modules handle the details of accessing some 50 different databases. There are DBD:: drivers for most ANSI SQL databases.

Comparative performance

The "Computer Language Shootout Benchmarks" compare the performance of implementations of typical programming problems in several programming languages. Their Perl implementations typically took up more memory than implementations in other languages, and had varied speed results. Perl's performance in the shootout is similar to other interpreted languages such as Python, PHP and Ruby, but slower than most compiled languages.

Perl can be slower than other languages doing the same thing because it has to compile the source every time it runs. In "A Timely Start", Jean-Louis Leroy found that his Perl scripts took much longer to run than he expected because the perl interpreter spent much of the time finding and compiling modules. Since most Perl programmers do not know how to save its intermediate result as Java, Python, and Ruby do easily, Perl scripts pay this overhead penalty on every execution. The overhead is not such a problem when amortized over a long run phase, but can significantly skew measurement of very short execution times as often found in benchmarks. Once perl starts the run phase, however, it can be quite fast and will typically outperform other dynamic languages. Technologies such as mod_perl overcome this by holding the compiled program in memory between multiple runs, or Class::Autouse to delay compiling of parts of the program until needed.

Optimizing

Nicholas Clark, a Perl core developer, discusses some Perl design trade-offs and some solutions in "When perl is not quite fast enough". The most critical routines of a Perl program can be written in other languages such as C or Assembler via XS or Inline.

Optimizing Perl can require intimate knowledge of its workings rather than skill with the language and its syntax, meaning that the problem is with the implementation of Perl rather than the language itself. Perl 6, the next major version, will address some of these lessons that other languages have already learned.

Future

At the 2000 Perl Conference, Jon Orwant made a case for a major new language initiative. This led to a decision to begin work on a redesign of the language, to be called Perl 6. Proposals for new language features were solicited from the Perl community at large, and over 300 RFCs were submitted.

Larry Wall spent the next few years digesting the RFCs and synthesizing them into a coherent framework for Perl 6. He has presented his design for Perl 6 in a series of documents called apocalypses, which are numbered to correspond to chapters in Programming Perl ("The Camel Book"). The current, unfinalized specification of Perl 6 is encapsulated in design documents called Synopses, which are numbered to correspond to Apocalypses.

Perl 6 is not intended to be backward compatible, though there will be a compatibility mode.

In 2001, it was decided that Perl 6 would run on a cross-language virtual machine called Parrot. This will mean that other languages targeting the Parrot will gain native access to CPAN and will allow some level of cross-language development.

In 2005 Audrey Tang created the pugs project, an implementation of Perl 6 in Haskell. This was and continues to act as a test platform for the Perl 6 language (separate from the development of the actual implementation) allowing the language designers to explore. The pugs project spawned an active Perl/Haskell cross-language community centered around the Freenode #perl6 irc channel.

A number of features in the Perl 6 language now show similarities with Haskell, and Perl 6 has been embraced by the Haskell community as a potential scripting language.

As of 2006, Perl 6, Parrot, and pugs are under active development, and a new module for Perl 5 called v6 allows some Perl 6 code to run directly on top of Perl 5.

In 2006, an effort was started to have Windows Perl distributions ship with a compiler, in order to make the need for binary packages on Windows redundant. Some early results of this include the CamelPack macro-installer and Vanilla Perl distributions.

Fun with Perl

Perl has a strong culture with many traditions, several of which are practiced purely for recreational value.

As with C, obfuscated code competitions are the most well-known pastime. The annual Obfuscated Perl contest made an arch virtue of Perl's syntactic flexibility. The following program prints the text "Just another Perl / Unix hacker", using 32 concurrent processes coordinated by pipes. A complete explanation is available on the author's Web site.

 @P=split//,".URRUU\c8R";@d=split//,"\nrekcah xinU / lreP rehtona tsuJ";sub p{
 @p{"r$p","u$p"}=(P,P);pipe"r$p","u$p";++$p;($q*=2)+=$f=!fork;map{$P=$P[$f^ord
 ($p{$_})&6];$p{$_}=/ ^$P/ix?$P:close$_}keys%p}p;p;p;p;p;map{$p{$_}=~/^[P.]/&&
 close$_}%p;wait until$?;map{/^r/&&<$_>}%p;$_=$d[$q];sleep rand(2)if/\S/;print

This is also an example of a discipline similar to obfuscated code, but somewhat distinct from it, known as the "JAPH." In the parlance of Perl culture, Perl programmers are known as Perl hackers, and from this derives the practice of writing short programs to print out the phrase " Just another Perl hacker,". In the spirit of the original concept, these programs are moderately obfuscated and short enough to fit into the signature of an email or Usenet message. The "canonical" JAPH includes the comma at the end, although this is often omitted. Many variants on the theme have been created, eg. , which prints "Just Another Perl Pirate!".

Another popular diversion is "Perl Golf," which has the same goal as the physical sport: to reduce the number of strokes that it takes to complete a particular objective. In this context, "strokes" refers to keystrokes, rather than swings of a golf club. Objectives are narrowly defined non-trivial tasks, such as "scan an input string and return the longest palindrome that it contains." Participants try to outdo each other by writing solutions that require ever fewer characters of Perl source code.

Similar to obfuscated code and golf, but with a different purpose, Perl poetry is the practice of writing poems that can actually be compiled as legal (although generally non-sensical) Perl code. This hobby is more or less unique to Perl due to the large number of regular English words used in the language. New poems are regularly published in the Perl Monks site's Perl Poetry section. Part of Perl lore is Black Perl, an infamous example of Perl poetry.

There are also many examples of code written purely for entertainment on the CPAN. Examples include the module Lingua::Romana::Perligata , which allows writing programs in Latin. Upon execution of such a program, the module translates its source code into regular Perl and runs it.

The Perl community has set aside the " Acme" namespace for modules that are fun in nature (but its scope has widened to include exploratory or experimental code or any other module that is not meant to ever be used in production). Some of the Acme modules are deliberately implemented in amusing ways. Some examples:

• Acme::Bleach, one of the first modules in the Acme:: namespace, allows the program's source code to be "whitened" (i.e., all characters replaced with whitespace) and yet still work. This is an example of a source filter. There are also a number of other source filters in the Acme namespace.
• Acme::Hello simplifies the process of writing a "Hello, World!" program
• Acme::Currency allows you to change the "$" prefix for scalar variables to some other character
• Acme::ProgressBar is a purposefully horribly inefficient way to indicate progress for a task
• Acme::VerySign satirizes the widely-criticized VeriSign Site Finder service
• Acme::Don't implements the logical opposite of the do keyword: the don't keyword, which takes a block that it does not execute. (It should be noted that when using this, don't{ ... } does not do the same thing as do not { ... }. It doesn't not, either.)