Python Types and Objects


Explains different Python new-style objects:

  • what are the <type 'type'> and <type 'object'> types
  • how user defined classes and instances are related to each other and to built-in types
  • what are metaclasses

The system described is sometimes called the Python type system, or the object model.

Read The Book


This book is part of a series:

  1. Python Types and Objects [you are here]

  2. Python Attributes and Methods


28 Jan, 2006

No-one commented on this great tutorial yet? For shame!

This (along with "Python Attributes and Methods") is the most lucid explanation of Python's OO system that i know of.

21 Dec, 2006

Your site has very much liked me. I shall necessarily tell about him to the friends.

11 Jan, 2007

Sehr guten site. Alles arbeitet deutlich(klar), schon eben storungsfrei. Wer machte? Vielleicht vom Weg?

14 Jan, 2007

Interesting, but navigation system is a little bit confusing

12 Apr, 2007

Excellent tutorial! This information will be very helpful for anyone exploring the Python C API.

17 Jun, 2007
Eiwot Towie:

Very good tutorial ! Thanks for your sharing. I also created Python articles at

23 Nov, 2007

Awesome! One of the best Python article. Thanks a ton!

07 Feb, 2008

Great article. You really know how to put it simple, clear and entertaining. I enjoyed every word of it.

08 May, 2008

Been looking all over for something like this. Well done.

10 Jun, 2008

Excellent tut! Ive been searching the net for these advanced topics. These 'books' are a MUST read for any pythonist that wants to dig up that python inner black magic.

05 Jul, 2008

Took me a while to read and grok them, but they're excellent and exactly what I was looking for. Thanks a bunch!

08 Aug, 2008

This cleared up a lot of my confusion... thanks! As an earlier post put it, "Your site has very much liked me. I shall necessarily tell about him to the friends."

09 Dec, 2008

This book is quite unique ! After more than 200 hours of reading and self-studying of Python (and I am still bad at it), this text is of tremendous quality, even for beginners ! It is concise ; It is very explicit and well illustrated; I have not seen a text so clear about the internals of Python for months;

Some people (e.g: me) need to know how a machine work before using it: you're the one that do it best for Python !

So for anyone who thinks that "objects" and "instances" are impalpable matter that flow between classes and methods, read the book !

24 Feb, 2009

Just want to thank you for an excellent summary of the fundamentals of objects in Python. I'm actually a beginner to Python but have been asking myself these sort of questions (of which you answer many in your chapters) about the fundamentals of how the language works. I feel like I've surpassed many so-called experienced Python programmers who never bother with these issues and just want to get a job done. Once again, thanks a lot for this great resource.

01 Mar, 2009

amazingly hard to understand.

13 May, 2009

Thanks for the great tutorial. Using the word class for "type" would simplify the discussion a lot. Design behind creating the circular reltationship between type and class is not explained here. That would have really helped.

02 Jul, 2009

I loved this site and have learned a lot. I could use python fairly well but had some confusing points. You cleared by confusion. Are you going to write article for other parts of python? Having deep understanding/knowledge on something is one thing and conveying this to others is another. You have both. THANKS

23 Aug, 2009

Thanks from Czech Rupublic.

26 Aug, 2009

I'm quite surprised at all the praise. Coming from a C++ background, I find the "type is an object" paradigm unnecessary, but I accept it has advantages. This article ("book"? please, curb your ego) has helped me understand the situation, but not without a lot of rereading.

The first source of confusion for me is where the word "base" is introduced with no mention that it describes the subtyping relationship.

The article shows isinstance(object, type) is True but doesn't also show that isinstance(type, object) is True The solid arrow connecting 'object' and 'type' in the diagram should be double-headed.

I think the "Can skim section" note is wrong: I do understand type/subtype and object/instance and I still had to reread that section several times, and then return to it after reading the later discussion, to understand what you were talking about. "Subtype an instance"? From a C++ perspective, that makes no sense.

27 Aug, 2009

The solid arrow is the issubclass relationship. The dashed arrow is the isinstance relationship. The dashed arrow could be double ended, but that is not necessary because it automatically follows from the other relationships (Example 2.2).

While subclassing an instance makes no sense in C++, in Python all classes are techincally instances and so subclassing makes sense.

If coming from C++, you basically have to forget the C++ concept of classes and assume all Python classes, meta-classes and objects are like C++ instances. All these 'instances' are located in memory and hold references to each other. There are two kinds of references (or edges in this graph of objects) - isinstance and issubclass. How the objects are connected determine how they behave.

Anyway, parts of this book seem too pedantic now and can probably be reduced or simplified.

09 Nov, 2009

In the chapter 3 Wrap Up Section Summary. issubtype maybe a typo of issubclass or not?

10 Nov, 2009

thank you so much for your Python Types and Objects online book - it has clarified so much that I have been struggling with, unable to get a clear answer. your flow chart maps are very helpful

11 Nov, 2009

Very fine analysis and explanation. Thanks for the tutorial. If the basics are cleared then anybody can even read the python source for the libraries and understand them easily as python is that fine. Try to put your book into the Python documentation section. It will be very useful for all the eyes that watch the docs. Oh and i forgot to ask 'How is Squasher?' ;)

21 Dec, 2009

this is crap

05 Feb, 2010

the commentary can be a little too cheesy at points...

23 Feb, 2010

Beautifully explained. Thank you!

14 Apr, 2010

Very Nice explanations ... Thanks you very much .. I was having a great problem with those concepts, but NOW, they are almost vanished >> Again ,, thanks alot .. You are really a python master

18 Apr, 2010

Just great! Basics for every python programmer.

07 May, 2010

This is fantastic. It made way more sense than any other explanation of any aspect of Python I've ever read. Very well written and excellently pitched. Maybe not for everyone but perfect for me.

11 May, 2010

Great explanation! Thank you for a wonderfully written tutorial.

10 Oct, 2010

A little bewildering, thanks god for the diagrams. But this is the only undertandable explanation of these things (maybe because I'm a visual thinker and I say your drawings). Some things about subclassing and metaclasses I don't quite get, but I'm not so deep in advanced python to need them anyway...


21 Nov, 2010

Thank you for the article. I am a newbie to Python. This is the first explanation of the language definition that seems to remove\explain the circular definitions. I suspect I will have to re-read later to understand it all but this really helps.

01 Dec, 2010

Excellent tutorial. Against your recommendations I tried it a few years ago when I was new to Python and understood nothing, but now with more experience, it made very much sense (more so than any other source I have tried).

two notes:

  • In python 3, type(list), type(object) etc now returns (with home-made html escapes, I can't get them to work)

lt class 'type' gt

rather than

lt type 'type' gt

I found that confusing at first -- you may wish to add a line about the change.

  • The following is now a fact rather than an assumption:

    ClassType object exists for backward compatibility and may not exist in future versions of Python


17 Dec, 2010

I would sum it up as follows: consider types, objects, classes and instances.

rule 1: Everything is an object (types are objects, objects are objects, classes are objects and instances are objects)

rule 2: Everything has a type (objects have a type, types have a type, classes have a type, and instances have a type)

Very nice tutorial and helps clarify a lot of otherwise muddy concepts.

20 Dec, 2010

a very intuitive way to explain the relationships between "type" and "class" in python. a great article ! thanks

06 Feb, 2011

What does type(NewStyleClass) return (class with the object class as a parent)? Does the enumeration value exist in the types module? What version of python does this enumeration value exist in?

  • Confused
07 May, 2011

Somebody translated the entire text into Chinese!

09 May, 2011

+++They can be subclassed. This means you can create a new object that is somewhat similar to existing type objects. The existing type objects become bases for the new one. ---They can be subclassed. This means you can create a new object that is somewhat similar to exsiting type objects. The existing type objects become bases for the new one.

02 Jun, 2011

Excellent guide, thanks. My only quibble is with this passage:

...if one is asked to list words similar to 'snake', one is likely to come up with 'reptile'. However, when asked to list words similar to 'Squasher', one is unlikely to say 'snake'.

Isn't the distinction here that "snake" and "reptile" are words in common use, whereas the proper name "Squasher" is not? Presumably, if one were familiar with Squasher the python, one would quickly think to list the word snake, the same way most everyone would say "dog" if asked to list words similar to "Snoopy".

07 Sep, 2011

A great book i ever came across. Thanks Shalabh ji . A great work Sir !!!

24 Dec, 2011

WOW! Eager to read what's in the next chapter.

Thanks a ton! I really appreciate it.

PS: ist hard to get captcha right. hoof!

12 Jan, 2012

I am so glad I came across this. It just nails the concepts perfectly.

Many thanks!

12 Jan, 2012

Excellent article.

I think there's a small bug with Example 2.2. I think that numbers 1 and 2 do not refer to the commands.

In general, it would be nice to try out, and explain, all four possibilities of isinstance with object and type.

Thank you for the great article.

16 Feb, 2012

This is brilliant!Iv always been stuck between "working with python" and "understanding its internals". To really understand the type system, the pep draft is pretty terse and its easy to feel lost. However this article brilliantly explains the theory of the object oriented type system of python, keeping the abstract spirit alive and not falling back on lame analogies and examples as any "for dummies" book would do.. Clearly this article is the stepping stone to all those terse and dense pep drafts that i can now face with some degree of confidence. Thanks!!! Wish u could do more of these kind of article on other areas of python under the hood

28 Feb, 2012

This is the most valuable tutorial. AWESOME!!!

03 Jul, 2012

nice tutorial

15 Aug, 2012

As Anonymous mentioned there is a small bug in Example 2.2 (part 1. should read isinstance(object, type) rather than isinstance(object, object).

Thanks for the article.

12 Sep, 2012

I don't think it's a bug, it's rather a feature. The text says: "instances of [HTML_REMOVED] are instances of [HTML_REMOVED] as well." As [HTML_REMOVED] is an instance of [HTML_REMOVED], it is an instance of [HTML_REMOVED] as well.

A very concise and helpful tutorial, thanks a lot. The first time that I started understanding how Python classes work. :)

12 Oct, 2012


Finally reached Enlightment in Python type/object hierarchy and I owe in to this article. This is simply an AWESOME tutorial, thanks a lot!!!!!!!!!!!!!!!!!!

27 Oct, 2012

Excellent tutorial, thank you so much for these books. Only flaws would be the occasional spelling/syntax mistakes (in English, not in Python).

03 Apr, 2013

Hi, Thanks for the share. But I think it would be great if you could place module into that Slate.

23 May, 2013

Very enlightening article! I think you may have made a mistake in Chapter 3 though:

issubclass(A,B) (testing for superclass-subclass relationship) returns True iff:
    1. B is in A.__bases__, or
    2. issubclass(Z,B) is true for any Z in A.__bases__.

This would mean issubclass(object, object) should return False, but (in Python 2.7) it returns True.

If change this, the explanation for isinstance can also be simplified.

23 Aug, 2013

The best Article. Cleared all my doubts.

May I ask for some more articles on topics like decorators, coroutines and context managers. Your suggestions for any links on above topics would be very helpful for me.

Thank you so much for sharing the best

30 Sep, 2013

Mistypes: ... are both objects but only types can have subcasses. -> subclasses

05 Nov, 2013

Typo - Chapter 2, the first bullet-point defining: type objects

Is written ...

They are used to represent abstract data types in programs. For instance, one (user defined) object called User might represent all users in a system, another once called int might represent all integers.

Should be written ...

They are used to represent abstract data types in programs. For instance, one (user defined) object called User might represent all users in a system, another one called int might represent all integers.

18 Jan, 2014

This is perhaps the best article i have read on the internet about python OO internals. Amazing!!

I did find something a bit confusing: """ Note that by just subclassing type 'object', the type C automatically is an instance of type 'type' """

From the diagram, It can be seen that 'object' is an instance of 'type' (only dotted line from 'object' to 'type'). Since there is no subclassing between them, i don't see how this is the reason isinstance(C, type) evaluates to TRUE.

I think (not entirely sure here) that C is an instance of 'type' because 'type' is used to instantiate C. And that type objects can only be instantiated through 'type'. Is this not the case in python? In the text, this case seems to be covered under the "Dashed Arrow Down Rule". But this rule would only hold for type objects. That is, the only way A can subclass B is if A itself is a type object, which would be instantiated by 'type' in any case, making isinstance(A, type) -> TRUE. So another question is that can type objects other than 'type' instantiate other type objects without relying on 'type'?

Again, thanks for this excellent gem.

06 May, 2014

Great article!

16 May, 2014
Justin H*:

I'm brand new to programming, and only about half-way through "teaching myself" Python.

That being said, I'll be honest, this took me a couple days. I've had to read this document a couple times, and I've had to stare at and contemplate a couple parts that I didn't quite "get" at first. However, now I feel a nice sense of accomplishment that everything's finally begun to "click" in my mind and come together.

I feel extremely satisfied with how thoroughly I understand this subject now, from the full picture down to the small details. I learned stuff I didn't even expect to learn. Resources like this that explain the full picture yet do it simply and coherently enough for a lowly beginner can be hard to come by. This is perfect for anyone who wants to be great at Python.

28 Aug, 2014

Someone above suggested including 'module' on the slate, and I agree, because: an instance of a module has no bases attribute. Why?

28 Aug, 2014

Too bad comments can't be deleted, because my previous one is pretty idiotic. :/

Post Comment
Sign In or provide:
Not disclosed
Human Test*
Markdown formatting
powered by durusworks