4 Consistent formatting of this file is achieved with:
5 perl ./Porting/podtidy pod/perlootut.pod
9 perlootut - Object-Oriented Programming in Perl Tutorial
13 This document was created in February, 2011, and the last major
14 revision was in February, 2013.
16 If you are reading this in the future then it's possible that the state
17 of the art has changed. We recommend you start by reading the perlootut
18 document in the latest stable release of Perl, rather than this
23 This document provides an introduction to object-oriented programming
24 in Perl. It begins with a brief overview of the concepts behind object
25 oriented design. Then it introduces several different OO systems from
26 L<CPAN|https://www.cpan.org> which build on top of what Perl
29 By default, Perl's built-in OO system is very minimal, leaving you to
30 do most of the work. This minimalism made a lot of sense in 1994, but
31 in the years since Perl 5.0 we've seen a number of common patterns
32 emerge in Perl OO. Fortunately, Perl's flexibility has allowed a rich
33 ecosystem of Perl OO systems to flourish.
35 If you want to know how Perl OO works under the hood, the L<perlobj>
36 document explains the nitty gritty details.
38 This document assumes that you already understand the basics of Perl
39 syntax, variable types, operators, and subroutine calls. If you don't
40 understand these concepts yet, please read L<perlintro> first. You
41 should also read the L<perlsyn>, L<perlop>, and L<perlsub> documents.
43 =head1 OBJECT-ORIENTED FUNDAMENTALS
45 Most object systems share a number of common concepts. You've probably
46 heard terms like "class", "object, "method", and "attribute" before.
47 Understanding the concepts will make it much easier to read and write
48 object-oriented code. If you're already familiar with these terms, you
49 should still skim this section, since it explains each concept in terms
50 of Perl's OO implementation.
52 Perl's OO system is class-based. Class-based OO is fairly common. It's
53 used by Java, C++, C#, Python, Ruby, and many other languages. There
54 are other object orientation paradigms as well. JavaScript is the most
55 popular language to use another paradigm. JavaScript's OO system is
60 An B<object> is a data structure that bundles together data and
61 subroutines which operate on that data. An object's data is called
62 B<attributes>, and its subroutines are called B<methods>. An object can
63 be thought of as a noun (a person, a web service, a computer).
65 An object represents a single discrete thing. For example, an object
66 might represent a file. The attributes for a file object might include
67 its path, content, and last modification time. If we created an object
68 to represent F</etc/hostname> on a machine named "foo.example.com",
69 that object's path would be "/etc/hostname", its content would be
70 "foo\n", and it's last modification time would be 1304974868 seconds
71 since the beginning of the epoch.
73 The methods associated with a file might include C<rename()> and
76 In Perl most objects are hashes, but the OO systems we recommend keep
77 you from having to worry about this. In practice, it's best to consider
78 an object's internal data structure opaque.
82 A B<class> defines the behavior of a category of objects. A class is a
83 name for a category (like "File"), and a class also defines the
84 behavior of objects in that category.
86 All objects belong to a specific class. For example, our
87 F</etc/hostname> object belongs to the C<File> class. When we want to
88 create a specific object, we start with its class, and B<construct> or
89 B<instantiate> an object. A specific object is often referred to as an
90 B<instance> of a class.
92 In Perl, any package can be a class. The difference between a package
93 which is a class and one which isn't is based on how the package is
94 used. Here's our "class declaration" for the C<File> class:
98 In Perl, there is no special keyword for constructing an object.
99 However, most OO modules on CPAN use a method named C<new()> to
100 construct a new object:
102 my $hostname = File->new(
103 path => '/etc/hostname',
105 last_mod_time => 1304974868,
108 (Don't worry about that C<< -> >> operator, it will be explained
113 As we said earlier, most Perl objects are hashes, but an object can be
114 an instance of any Perl data type (scalar, array, etc.). Turning a
115 plain data structure into an object is done by B<blessing> that data
116 structure using Perl's C<bless> function.
118 While we strongly suggest you don't build your objects from scratch,
119 you should know the term B<bless>. A B<blessed> data structure (aka "a
120 referent") is an object. We sometimes say that an object has been
121 "blessed into a class".
123 Once a referent has been blessed, the C<blessed> function from the
124 L<Scalar::Util> core module can tell us its class name. This subroutine
125 returns an object's class when passed an object, and false otherwise.
127 use Scalar::Util 'blessed';
129 print blessed($hash); # undef
130 print blessed($hostname); # File
134 A B<constructor> creates a new object. In Perl, a class's constructor
135 is just another method, unlike some other languages, which provide
136 syntax for constructors. Most Perl classes use C<new> as the name for
139 my $file = File->new(...);
143 You already learned that a B<method> is a subroutine that operates on
144 an object. You can think of a method as the things that an object can
145 I<do>. If an object is a noun, then methods are its verbs (save, print,
148 In Perl, methods are simply subroutines that live in a class's package.
149 Methods are always written to receive the object as their first
155 print "This file is at ", $self->path, "\n";
159 # The file is at /etc/hostname
161 What makes a method special is I<how it's called>. The arrow operator
162 (C<< -> >>) tells Perl that we are calling a method.
164 When we make a method call, Perl arranges for the method's B<invocant>
165 to be passed as the first argument. B<Invocant> is a fancy name for the
166 thing on the left side of the arrow. The invocant can either be a class
167 name or an object. We can also pass additional arguments to the method:
171 my $prefix = shift // "This file is at ";
173 print $prefix, ", ", $self->path, "\n";
176 $file->print_info("The file is located at ");
177 # The file is located at /etc/hostname
181 Each class can define its B<attributes>. When we instantiate an object,
182 we assign values to those attributes. For example, every C<File> object
183 has a path. Attributes are sometimes called B<properties>.
185 Perl has no special syntax for attributes. Under the hood, attributes
186 are often stored as keys in the object's underlying hash, but don't
189 We recommend that you only access attributes via B<accessor> methods.
190 These are methods that can get or set the value of each attribute. We
191 saw this earlier in the C<print_info()> example, which calls C<<
194 You might also see the terms B<getter> and B<setter>. These are two
195 types of accessors. A getter gets the attribute's value, while a setter
196 sets it. Another term for a setter is B<mutator>
198 Attributes are typically defined as read-only or read-write. Read-only
199 attributes can only be set when the object is first created, while
200 read-write attributes can be altered at any time.
202 The value of an attribute may itself be another object. For example,
203 instead of returning its last mod time as a number, the C<File> class
204 could return a L<DateTime> object representing that value.
206 It's possible to have a class that does not expose any publicly
207 settable attributes. Not every class has attributes and methods.
211 B<Polymorphism> is a fancy way of saying that objects from two
212 different classes share an API. For example, we could have C<File> and
213 C<WebPage> classes which both have a C<print_content()> method. This
214 method might produce different output for each class, but they share a
217 While the two classes may differ in many ways, when it comes to the
218 C<print_content()> method, they are the same. This means that we can
219 try to call the C<print_content()> method on an object of either class,
220 and B<we don't have to know what class the object belongs to!>
222 Polymorphism is one of the key concepts of object-oriented design.
226 B<Inheritance> lets you create a specialized version of an existing
227 class. Inheritance lets the new class reuse the methods and attributes
230 For example, we could create an C<File::MP3> class which B<inherits>
231 from C<File>. An C<File::MP3> B<is-a> I<more specific> type of C<File>.
232 All mp3 files are files, but not all files are mp3 files.
234 We often refer to inheritance relationships as B<parent-child> or
235 C<superclass>/C<subclass> relationships. Sometimes we say that the
236 child has an B<is-a> relationship with its parent class.
238 C<File> is a B<superclass> of C<File::MP3>, and C<File::MP3> is a
239 B<subclass> of C<File>.
245 The L<parent> module is one of several ways that Perl lets you define
246 inheritance relationships.
248 Perl allows multiple inheritance, which means that a class can inherit
249 from multiple parents. While this is possible, we strongly recommend
250 against it. Generally, you can use B<roles> to do everything you can do
251 with multiple inheritance, but in a cleaner way.
253 Note that there's nothing wrong with defining multiple subclasses of a
254 given class. This is both common and safe. For example, we might define
255 C<File::MP3::FixedBitrate> and C<File::MP3::VariableBitrate> classes to
256 distinguish between different types of mp3 file.
258 =head3 Overriding methods and method resolution
260 Inheritance allows two classes to share code. By default, every method
261 in the parent class is also available in the child. The child can
262 explicitly B<override> a parent's method to provide its own
263 implementation. For example, if we have an C<File::MP3> object, it has
264 the C<print_info()> method from C<File>:
266 my $cage = File::MP3->new(
267 path => 'mp3s/My-Body-Is-a-Cage.mp3',
268 content => $mp3_data,
269 last_mod_time => 1304974868,
270 title => 'My Body Is a Cage',
274 # The file is at mp3s/My-Body-Is-a-Cage.mp3
276 If we wanted to include the mp3's title in the greeting, we could
286 print "This file is at ", $self->path, "\n";
287 print "Its title is ", $self->title, "\n";
291 # The file is at mp3s/My-Body-Is-a-Cage.mp3
292 # Its title is My Body Is a Cage
294 The process of determining what method should be used is called
295 B<method resolution>. What Perl does is look at the object's class
296 first (C<File::MP3> in this case). If that class defines the method,
297 then that class's version of the method is called. If not, Perl looks
298 at each parent class in turn. For C<File::MP3>, its only parent is
299 C<File>. If C<File::MP3> does not define the method, but C<File> does,
300 then Perl calls the method in C<File>.
302 If C<File> inherited from C<DataSource>, which inherited from C<Thing>,
303 then Perl would keep looking "up the chain" if necessary.
305 It is possible to explicitly call a parent method from a child:
314 $self->SUPER::print_info();
315 print "Its title is ", $self->title, "\n";
318 The C<SUPER::> bit tells Perl to look for the C<print_info()> in the
319 C<File::MP3> class's inheritance chain. When it finds the parent class
320 that implements this method, the method is called.
322 We mentioned multiple inheritance earlier. The main problem with
323 multiple inheritance is that it greatly complicates method resolution.
324 See L<perlobj> for more details.
328 B<Encapsulation> is the idea that an object is opaque. When another
329 developer uses your class, they don't need to know I<how> it is
330 implemented, they just need to know I<what> it does.
332 Encapsulation is important for several reasons. First, it allows you to
333 separate the public API from the private implementation. This means you
334 can change that implementation without breaking the API.
336 Second, when classes are well encapsulated, they become easier to
337 subclass. Ideally, a subclass uses the same APIs to access object data
338 that its parent class uses. In reality, subclassing sometimes involves
339 violating encapsulation, but a good API can minimize the need to do
342 We mentioned earlier that most Perl objects are implemented as hashes
343 under the hood. The principle of encapsulation tells us that we should
344 not rely on this. Instead, we should use accessor methods to access the
345 data in that hash. The object systems that we recommend below all
346 automate the generation of accessor methods. If you use one of them,
347 you should never have to access the object as a hash directly.
351 In object-oriented code, we often find that one object references
352 another object. This is called B<composition>, or a B<has-a>
355 Earlier, we mentioned that the C<File> class's C<last_mod_time>
356 accessor could return a L<DateTime> object. This is a perfect example
357 of composition. We could go even further, and make the C<path> and
358 C<content> accessors return objects as well. The C<File> class would
359 then be B<composed> of several other objects.
363 B<Roles> are something that a class I<does>, rather than something that
364 it I<is>. Roles are relatively new to Perl, but have become rather
365 popular. Roles are B<applied> to classes. Sometimes we say that classes
368 Roles are an alternative to inheritance for providing polymorphism.
369 Let's assume we have two classes, C<Radio> and C<Computer>. Both of
370 these things have on/off switches. We want to model that in our class
373 We could have both classes inherit from a common parent, like
374 C<Machine>, but not all machines have on/off switches. We could create
375 a parent class called C<HasOnOffSwitch>, but that is very artificial.
376 Radios and computers are not specializations of this parent. This
377 parent is really a rather ridiculous creation.
379 This is where roles come in. It makes a lot of sense to create a
380 C<HasOnOffSwitch> role and apply it to both classes. This role would
381 define a known API like providing C<turn_on()> and C<turn_off()>
384 Perl does not have any built-in way to express roles. In the past,
385 people just bit the bullet and used multiple inheritance. Nowadays,
386 there are several good choices on CPAN for using roles.
388 =head2 When to Use OO
390 Object Orientation is not the best solution to every problem. In I<Perl
391 Best Practices> (copyright 2004, Published by O'Reilly Media, Inc.),
392 Damian Conway provides a list of criteria to use when deciding if OO is
393 the right fit for your problem:
399 The system being designed is large, or is likely to become large.
403 The data can be aggregated into obvious structures, especially if
404 there's a large amount of data in each aggregate.
408 The various types of data aggregate form a natural hierarchy that
409 facilitates the use of inheritance and polymorphism.
413 You have a piece of data on which many different operations are
418 You need to perform the same general operations on related types of
419 data, but with slight variations depending on the specific type of data
420 the operations are applied to.
424 It's likely you'll have to add new data types later.
428 The typical interactions between pieces of data are best represented by
433 The implementation of individual components of the system is likely to
438 The system design is already object-oriented.
442 Large numbers of other programmers will be using your code modules.
446 =head1 PERL OO SYSTEMS
448 As we mentioned before, Perl's built-in OO system is very minimal, but
449 also quite flexible. Over the years, many people have developed systems
450 which build on top of Perl's built-in system to provide more features
453 We strongly recommend that you use one of these systems. Even the most
454 minimal of them eliminates a lot of repetitive boilerplate. There's
455 really no good reason to write your classes from scratch in Perl.
457 If you are interested in the guts underlying these systems, check out
462 L<Moose> bills itself as a "postmodern object system for Perl 5". Don't
463 be scared, the "postmodern" label is a callback to Larry's description
464 of Perl as "the first postmodern computer language".
466 C<Moose> provides a complete, modern OO system. Its biggest influence
467 is the Common Lisp Object System, but it also borrows ideas from
468 Smalltalk and several other languages. C<Moose> was created by Stevan
469 Little, and draws heavily from his work on the Raku OO design.
471 Here is our C<File> class using C<Moose>:
476 has path => ( is => 'ro' );
477 has content => ( is => 'ro' );
478 has last_mod_time => ( is => 'ro' );
483 print "This file is at ", $self->path, "\n";
486 C<Moose> provides a number of features:
490 =item * Declarative sugar
492 C<Moose> provides a layer of declarative "sugar" for defining classes.
493 That sugar is just a set of exported functions that make declaring how
494 your class works simpler and more palatable. This lets you describe
495 I<what> your class is, rather than having to tell Perl I<how> to
496 implement your class.
498 The C<has()> subroutine declares an attribute, and C<Moose>
499 automatically creates accessors for these attributes. It also takes
500 care of creating a C<new()> method for you. This constructor knows
501 about the attributes you declared, so you can set them when creating a
504 =item * Roles built-in
506 C<Moose> lets you define roles the same way you define classes:
508 package HasOnOffSwitch;
526 =item * A miniature type system
528 In the example above, you can see that we passed C<< isa => 'Bool' >>
529 to C<has()> when creating our C<is_on> attribute. This tells C<Moose>
530 that this attribute must be a boolean value. If we try to set it to an
531 invalid value, our code will throw an error.
533 =item * Full introspection and manipulation
535 Perl's built-in introspection features are fairly minimal. C<Moose>
536 builds on top of them and creates a full introspection layer for your
537 classes. This lets you ask questions like "what methods does the File
538 class implement?" It also lets you modify your classes
541 =item * Self-hosted and extensible
543 C<Moose> describes itself using its own introspection API. Besides
544 being a cool trick, this means that you can extend C<Moose> using
547 =item * Rich ecosystem
549 There is a rich ecosystem of C<Moose> extensions on CPAN under the
550 L<MooseX|https://metacpan.org/search?q=MooseX>
551 namespace. In addition, many modules on CPAN already use C<Moose>,
552 providing you with lots of examples to learn from.
554 =item * Many more features
556 C<Moose> is a very powerful tool, and we can't cover all of its
557 features here. We encourage you to learn more by reading the C<Moose>
558 documentation, starting with
559 L<Moose::Manual|https://metacpan.org/pod/Moose::Manual>.
563 Of course, C<Moose> isn't perfect.
565 C<Moose> can make your code slower to load. C<Moose> itself is not
566 small, and it does a I<lot> of code generation when you define your
567 class. This code generation means that your runtime code is as fast as
568 it can be, but you pay for this when your modules are first loaded.
570 This load time hit can be a problem when startup speed is important,
571 such as with a command-line script or a "plain vanilla" CGI script that
572 must be loaded each time it is executed.
574 Before you panic, know that many people do use C<Moose> for
575 command-line tools and other startup-sensitive code. We encourage you
576 to try C<Moose> out first before worrying about startup speed.
578 C<Moose> also has several dependencies on other modules. Most of these
579 are small stand-alone modules, a number of which have been spun off
580 from C<Moose>. C<Moose> itself, and some of its dependencies, require a
581 compiler. If you need to install your software on a system without a
582 compiler, or if having I<any> dependencies is a problem, then C<Moose>
583 may not be right for you.
587 If you try C<Moose> and find that one of these issues is preventing you
588 from using C<Moose>, we encourage you to consider L<Moo> next. C<Moo>
589 implements a subset of C<Moose>'s functionality in a simpler package.
590 For most features that it does implement, the end-user API is
591 I<identical> to C<Moose>, meaning you can switch from C<Moo> to
592 C<Moose> quite easily.
594 C<Moo> does not implement most of C<Moose>'s introspection API, so it's
595 often faster when loading your modules. Additionally, none of its
596 dependencies require XS, so it can be installed on machines without a
599 One of C<Moo>'s most compelling features is its interoperability with
600 C<Moose>. When someone tries to use C<Moose>'s introspection API on a
601 C<Moo> class or role, it is transparently inflated into a C<Moose>
602 class or role. This makes it easier to incorporate C<Moo>-using code
603 into a C<Moose> code base and vice versa.
605 For example, a C<Moose> class can subclass a C<Moo> class using
606 C<extends> or consume a C<Moo> role using C<with>.
608 The C<Moose> authors hope that one day C<Moo> can be made obsolete by
609 improving C<Moose> enough, but for now it provides a worthwhile
610 alternative to C<Moose>.
612 =head2 Class::Accessor
614 L<Class::Accessor> is the polar opposite of C<Moose>. It provides very
615 few features, nor is it self-hosting.
617 It is, however, very simple, pure Perl, and it has no non-core
618 dependencies. It also provides a "Moose-like" API on demand for the
619 features it supports.
621 Even though it doesn't do much, it is still preferable to writing your
622 own classes from scratch.
624 Here's our C<File> class with C<Class::Accessor>:
627 use Class::Accessor 'antlers';
629 has path => ( is => 'ro' );
630 has content => ( is => 'ro' );
631 has last_mod_time => ( is => 'ro' );
636 print "This file is at ", $self->path, "\n";
639 The C<antlers> import flag tells C<Class::Accessor> that you want to
640 define your attributes using C<Moose>-like syntax. The only parameter
641 that you can pass to C<has> is C<is>. We recommend that you use this
642 Moose-like syntax if you choose C<Class::Accessor> since it means you
643 will have a smoother upgrade path if you later decide to move to
646 Like C<Moose>, C<Class::Accessor> generates accessor methods and a
647 constructor for your class.
651 Finally, we have L<Class::Tiny>. This module truly lives up to its
652 name. It has an incredibly minimal API and absolutely no dependencies
653 on any recent Perl. Still, we think it's a lot easier to use than
654 writing your own OO code from scratch.
656 Here's our C<File> class once more:
659 use Class::Tiny qw( path content last_mod_time );
664 print "This file is at ", $self->path, "\n";
669 With C<Class::Tiny>, all accessors are read-write. It generates a
670 constructor for you, as well as the accessors you define.
672 You can also use L<Class::Tiny::Antlers> for C<Moose>-like syntax.
676 As we mentioned before, roles provide an alternative to inheritance,
677 but Perl does not have any built-in role support. If you choose to use
678 Moose, it comes with a full-fledged role implementation. However, if
679 you use one of our other recommended OO modules, you can still use
680 roles with L<Role::Tiny>
682 C<Role::Tiny> provides some of the same features as Moose's role
683 system, but in a much smaller package. Most notably, it doesn't support
684 any sort of attribute declaration, so you have to do that by hand.
685 Still, it's useful, and works well with C<Class::Accessor> and
688 =head2 OO System Summary
690 Here's a brief recap of the options we covered:
696 C<Moose> is the maximal option. It has a lot of features, a big
697 ecosystem, and a thriving user base. We also covered L<Moo> briefly.
698 C<Moo> is C<Moose> lite, and a reasonable alternative when Moose
699 doesn't work for your application.
701 =item * L<Class::Accessor>
703 C<Class::Accessor> does a lot less than C<Moose>, and is a nice
704 alternative if you find C<Moose> overwhelming. It's been around a long
705 time and is well battle-tested. It also has a minimal C<Moose>
706 compatibility mode which makes moving from C<Class::Accessor> to
709 =item * L<Class::Tiny>
711 C<Class::Tiny> is the absolute minimal option. It has no dependencies,
712 and almost no syntax to learn. It's a good option for a super minimal
713 environment and for throwing something together quickly without having
714 to worry about details.
716 =item * L<Role::Tiny>
718 Use C<Role::Tiny> with C<Class::Accessor> or C<Class::Tiny> if you find
719 yourself considering multiple inheritance. If you go with C<Moose>, it
720 comes with its own role implementation.
724 =head2 Other OO Systems
726 There are literally dozens of other OO-related modules on CPAN besides
727 those covered here, and you're likely to run across one or more of them
728 if you work with other people's code.
730 In addition, plenty of code in the wild does all of its OO "by hand",
731 using just the Perl built-in OO features. If you need to maintain such
732 code, you should read L<perlobj> to understand exactly how Perl's
737 As we said before, Perl's minimal OO system has led to a profusion of
738 OO systems on CPAN. While you can still drop down to the bare metal and
739 write your classes by hand, there's really no reason to do that with
742 For small systems, L<Class::Tiny> and L<Class::Accessor> both provide
743 minimal object systems that take care of basic boilerplate for you.
745 For bigger projects, L<Moose> provides a rich set of features that will
746 let you focus on implementing your business logic. L<Moo> provides a
747 nice alternative to L<Moose> when you want a lot of features but need
748 faster compile time or to avoid XS.
750 We encourage you to play with and evaluate L<Moose>, L<Moo>,
751 L<Class::Accessor>, and L<Class::Tiny> to see which OO system is right