When I started learning
Python, I already knew some programming cause I had had to learn Java at university and I had already done some work in PHP. I was looking for something fun that would allow me to do non-webby things and someone, I can't remember who it was, recommended Python to me. I looked at the
website and docs and kinda got things going but I relatively quickly decided I wanted to learn things in a more structured way so I got myself a book. That book was
not the first Edition of Dive Into Python, it was some other book, I can't even remember its name. It had the usual dry examples that never felt "real" and it treated you like a dummy not giving you the feeling of real progress.
A few weeks later I got the first Dive into Python (you can read it for free on the
Dive Into Python website). I don't know how broken my mind already was from all the Java and PHP but Dive Into Python fixed it: Great writing style and real-world examples that you could apply to your day-to-day problems without
trying too hard. I have recommended Dive Into Python to numerous people asking about Python and everybody was really happy with it.
When Python3 (a newer version of Python that introduced a few backwards-incompatible changes to the language to clean it up and remove a few "warts that it had grown) the author
Mark Pilgrim (he works for Google on HTML5 and has an absolutely brilliant blog, he also
is on Twitter) decided to update the book to
Dive Into Python 3. A few months ago he asked on Twitter who wanted a review copy, I said yes and yesterday got a package from him containing a review copy of the book and a few nice words handwritten on the first page - to make a long story short [Disclaimer] I did not pay for the book I'm writing about, it was a gift by Mark, so take this with a grain of salt [/Disclaimer]. With all that said, let's go into detail.
When starting at the beginning with the first chapter I instantly got the impression that this book might actually work better for complete programming newbies than the first edition did. The old book was sometimes hard to understand if you didn't know how to program, you kinda needed to already know the basics and Dive Into Python supplied you with Python's syntax and more importantly a pythonic approach to solving real-world problems. But you had to bring something to the table. Dive Into Python 3 makes it easier to get into the whole topic, but I still wouldn't recommend it to a complete newbie: Knowing any other programming language (well maybe not Prolog) will enable you to quickly pickup what's going on and learn Python with this book. Being able to hack some PHP is enough. On the other hand this book does try to teach you a pythonic approach. There is a lot of bad Python code out there (often written by people with too much Java background) that is not idiomatic, that does things in a way you might use in Java but that's just not how you should use Python. So if you know a lot about some other programming platform, it might actually help to "forget" a little before reading the book to allow you to understand everything. That is not a weakness of the book, I just wanted to make sure who the target audience is.
I'll skip over the "howto install python on your machine" chapter because ... well ... a trained monkey can probably install software and there are more screenshots than there are in ... something really screenshotty. It's clear it's explicit but most people probably won't need that chapter.
Chapter 1 introduces you to Python. You get a real world example ("humanize a number" as in write a readable representation for X bytes as for example 1.3 TB). You get the full code and Mark walks you through how Python modules work, how functions are defined and how you call them. Simple straightforward. Throughout the book there are little info boxes that summarize important hints or gotchas which I found really helpful.
The next chapters deal with Python's native datatypes and how to work with them, but instead of just listing them and going through them, Mark first focusses on lists (lists are really important in Python, you do pretty much everything with lists) and list comprehensions. This allows him to do interesting things (like working with directories and files) at a rather early stage and people can write cool programs after having read the first 3 chapters.
Then he dives into the big one: Strings. Strings have changed a lot from Python2 to Python3 (since Python3 all strings are unicode) so he spends quite a lot of time in the beginning of the chapter explaining what strings actually are and why there are problems. I enjoyed the text but I can imagine that some people might actually not care too much about the underlying thoughts and problems. Maybe a quick note in the beginning of the textblock would have helped because it looks kinda scary considering how the book had been rather "light" with big text paragraphs before. But after that you just get a quick walk through how to deal with strings which kinda fits to the chapter following: Regular Expressions.
Many people think regular expressions are weird or scary and Mark doesn't even try to explain them completely. He builds the chapter around a case study (Street Addresses) and that works really well: Instead of throwing too much theory at the reader, the example helps you to get into the topic and with Mark holding your hand, you walk through the whole case study without any troubles. You will not be able to write huge regex all by yourself, but you know how they work and there is enough groundwork laid out for you to understand the official documentation (which is rather theoretical) a lot better.
The next chapters go through "generators and closures" with "classes and iterators" following. Which might look weird to you considering that Python is object-oriented, but allowing that kind of learning is actually one of the strengths of Python. It allows Mark to build on all the topics he covered earlier and allows him not to use interesting examples for classes (in this case he builds a fibonacci number iterator). What I missed in that chapter was an explanation of inheritance and how it works in Python. "X inherits Y" is used in the iterator explanation but I think that should have been made more explicit and it also should probably have gotten another example to show people when it might make sense to generalize some functionality into a superclass. If people have no idea how inheritance works that part will be quite a hard nut to crack I think.
Chapter 9 goes through unit testing. Unit testing is important but most people don't do it. Why? Cause it looks like a lot of extra work. Mark does build on a previous example of "Converting numbers to roman numerals" and shows how unit tests can really save your ass. Especially in connection to chapter 10 "refactoring" it's make clear why you want to use unittests and how to write them with the build-in unittest module. Those two chapters should be read by unittest-sceptics no matter what language they write code in. Great work here.
Chapters 11 to 14 ("Files", "XML", "Serializing Python Objects" and "HTTP Web Services") work together really well. You learn how to read data from files, how to deal with XML files (using the lxml package) and how to access HTTP apis. Those are real work problems and from the book it all flows really very naturally. Knowledge gained from reading data from files smoothly gets "upgraded" to knowledge about querying for example Flickr or Google apis. At this point the reader can solve many of his (or her) real-world problems with Python. The HTTP chapter is somewhat struggling with the fact that there is quite a lot about HTTP that you have to know (GET vs POST, redirecting, headers) so it does not really work as well as the other chapters: It ends with short code examples on how to post a dent to
Identi.ca but the chapter before that throws a lot of "basics" at the user that he (or she) doesn't really care that much about. Maybe having one clearer example would have helped here. It's not that the chapter is bad, every subsection in itself is clear and explains exactly why something is relevant and how it works, I just find the subsections don't tie into each other as they do in previous chapters.
For those amongst you that already know Python2, Chapter 15 is pure gold. Mark goes through the whole process of porting a real life Python2 library to Python3 explaining all the tools to support you Python3 comes with and where and how to fix stuff manually when automatic translation fails. Really helpful and cool chapter!
The last "real" chapter deals with something bad... packaging. If you know Python you also know that its packaging is a huge pain. "easy_install" doesn't integrate well with linux distributions' own package management facilities and it cannot uninstall which completely sucks. On the other hand I think it makes sense to tell people how to publish their code in a rather sane way instead of letting them fight PyPI (Python Package Index) and Distribute all alone. The chapter provides the user with a clear and simple walkthrough on how to package and publish python code. It suggests a directory layout and adds some more best practices so while Python packaging still sucks, I think that this chapter will still help people a lot.
The book finishes with 2 appendices, one showing you how to port code from Python2 to Python3 automatically and the other one listing the "magic functions" python objects have and what to do with them (good reference material).
All in all this is a great book if you want to get into Python. The examples given in most chapters are way better than what many programming books offer, they feel "real". Another great thing is the whole flow of the book: The book goes through aspects not based on some abstract categorization but based on what works together and what allows the user to use the knowledge gained in the previous chapters directly in the next one to do something cool. "Classes" is a good example: Instead of just throwing them at the user very early ("cause Python is object-oriented") it's kept simple for the user, purely based on functions and classes enter the mix only when we need them. The language (as in writing style) is very accessible and avoids unnecessary complexity, I think even people with English as a second language will not have trouble understanding the book.
The layout looks beautiful and the fonts are quite good, too, even though code sometimes "drowns" in text. Choosing somewhat more differing fonts might have been a good idea just for quick scanability of the page.
The
Website states that the book is:
The book is freely licensed under the Creative Commons Attribution Share-Alike license. You can download it as HTML or PDF. Über-geeks can also clone the Mercurial repository
. On the other hand the book says that
All rights reserved. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage or retrieval system, without the prior written permission of the copyright owner and the publisher.
I send an email to Mark, maybe he can clear that up, I guess the copyright is not about the content but the layout and stuff, but I'll keep you guys updated.
Dive Into Python 3 is available on
Amazon(affiliate link,
this one isn't
)
and most other bookstores so don't hesitate, it is a great book. And if you are unsure and want to check a chapter out or two (or the whole book), visit the
Dive Into Python 3 website
I got lucky enough to have received a 
