34
loading...
Let's Read – Eloquent Ruby – Ch 1
Perhaps my personal favorite recommendation for learning to program Ruby like a Rubyist, Eloquent Ruby is a book I recommend frequently to this day. That said, it was released in 2011 and things have changed a bit since then.
This series will focus on reading over Eloquent Ruby, noting things that may have changed or been updated since 2011 (around Ruby 1.9.2) to today (2021 — Ruby 3.0.x).
Without further ado, let’s get started!
Note: This is an updated version of a previous unfinished Medium series of mine you can find here.
We start in with an introduction on writing code that looks and feels like Ruby.
The first example in practice is the following class:
class Document
attr_accessor :title, :author, :content
def initialize(title, author, content)
@title = title
@author = author
@content = content
end
def words
@content.split
end
def word_count
words.size
end
endNow this code will look the much the same today, and Russ mentions right afterwards a style rule:
“The thing to note about the code above is that it follows the Ruby indentation convention: In Ruby you indent your code with two spaces per level” — Eloquent Ruby Ch 1
This style is still prevalent in the community, as mentioned in both the Ruby Style Guide and in the lint tool Rubocop. Note, that means spaces, not tabs, as Russ also mentioned:
“…so idiomatic Ruby should be serenely tab free” — Eloquent Ruby Ch 1
“The real questions regarding comments are when and how much” — Eloquent Ruby Ch 1
This is always a fun issue in code, the idea of self-documenting code versus explicit comments.
Russ encourages practical examples and how-to guides which are especially useful for people trying to figure out how to use your code later:
# Class that models a plain text document, complete with title
# and author:
#
# doc = Document.new('Hamlet', 'Shakespeare', 'To be or...')
# puts doc.title
# puts doc.author
# puts doc.content
#
# Document instances know how to parse their content into words:
#
# puts doc.words
# puts doc.word_countAs an aside, it’s rare to see spaces before or after parens in Ruby code in the wild, as mentioned in the Ruby Style Guide
Document.new( 'Hamlet', 'Shakespeare', 'To be or...' )Document.new('Hamlet', 'Shakespeare', 'To be or...')With more modern versions, I would highly encourage the use of YARDoc as a standard way of documenting your code. This establishes a formalism in your comments as well as a method of generating an entire documentation set for your users with a few commands.
I use YARDoc for Qo to document my code both inline and in a full documentation site that can be hosted on Github pages along with the repository.
So instead of the comment provided in the chapter, we might use YARDoc to produce something like this:
# Class that models a plain text document, complete with title
# and author.
#
# @example
#
# doc = Document.new('Hamlet', 'Shakespeare', 'To be or...')
# puts doc.title
# puts doc.author
# puts doc.content
#
# @author Russ
#
class Document
# Title of the document
attr_accessor :title
# Author of the document
attr_accessor :author
# Content of the document
attr_accessor :content
# Creates a new instance of a Document
#
# @param title [String]
# Title of the Document
#
# @param author [String]
# Author of the Document
#
# @param content [String]
# The content of the Document
#
# @return [Document]
def initialize(title, author, content)
@title = title
@author = author
@content = content
end
# Gets the individual words from our content
#
# @return [Array[String]]
def words
@content.split
end
# Gets the count of words in the document
#
# @return [Integer]
def word_count
words.size
end
endNow it may be overkill to comment on the individual attribute accessors for a class like this, but in some cases that may well help the developer using your code to understand what each attribute is used for.
How to comments are mentioned next as a method of explaining particularly complicated bits of code:
# Using ngram analysis, compute the probability
# that this document and the one passed in were
# written by the same person, This algorithm is
# known to be valid for American English and will
# probably work for British and Canadian English.
#
def same_author_probability(other_document)
# Implementation left as an exercise for the reader...
endPersonally I like these explanations for more complex bits of code, as is mentioned in the book. With YARDoc we can also add some nifty little external references using the
@see tag to point to something like Wikipedia or another resource for further reading:# @see https://en.wikipedia.org/wiki/N-gram N-Grams on WikipediaNow there are cases of inline comments, and when that might be appropriate or not:
return 0 if divisor == 0 # Avoid division by zero
count += 1 # Add one to countThe first one makes some sense, as it addresses a potential edge case. We could also note this in the documentation of such a method using a second return:
# Divides two numbers
#
# @param a [Number]
# @param b [Number]
#
# @return [Number]
# Result of dividing two numbers
#
# @return [0]
# Zero is returned if the divisor is zero, to prevent divide
# by zero errors.This can be seen as a matter of preference, but in this case it’s also exposed in later generated documentation as a top-level edge case that’s addressed by an early return.
The second, as is mentioned in the book, is rather silly. Russ addresses this:
“The danger in comments that explain how the code works is that they can easily slide off into the worst reason for adding comments: to make a badly written program somewhat comprehensible.” — Eloquent Ruby Ch 1
Ideally the names of classes, variables, and methods can be used to give clarity to the intent of the code so such comments are unnecessary.
That said, even the clearest of code benefits from a level of discoverability through good documentation in a standard format. Self-explanatory code is great, and I agree that it needs to be written in such a manner, but empowering new users through documentation is what separates an accessible library from a confounding one.
The challenge then becomes updating the documentation in sync with the actual code, because the only thing worse than no comments are misleading ones.
This section has stayed fairly consistent over the years in that…
Classes are camel-cased:
class FunDocumentConstants are screaming-snake-cased:
ANTLERS_PER_MALE_MOOSE = 2Variables are snake-cased:
lowercase_words_separated_by_underscoreThere was some contention noted on constant names between
CamelCase and SCREAMING_SNAKE_CASE, but over the past few years Rubyists have moved towards the latter as was Russ’s preference in the book.This one, amusingly, has become a bit of a regional issue. In Seattle parentheses are near universally avoided in a style called Seattle Ruby, unless absolutely necessary to not break the syntax parser.
The more commonly agreed upon items by Rubyists are:
Method arguments should be surrounded by parentheses:
def find_document(title, author)Methods with no arguments should omit parens on both definition and calling:
def words
@content.split
end
wordsConditionals should not use parens:
if (condition) # Bad
if condition # GoodParens are also typically avoided in single arity methods, or block style single arity methods:
puts "text"
describe "An RSpec test" do
# ...
endThough granted that
puts isn’t single arity:puts "Some", "words"I’ve seen that one go both ways, it’s a matter of preference by that point.
Ruby allows you to use semicolons
; to put quite a few things on one line. Absent code golfing and shell one-liners, this should be avoided except in a few cases as mentioned by the book:class DocumentException < Exception; end
def method_to_be_overridden; endNote though that in the second case if parens are present Ruby can tell the difference:
def add(a, b) a + b endThough whitespace is free, so use what makes sense and doesn’t give you lines that are ridiculously long just to avoid pressing return.
In Ruby 3.x, however, the one-liner syntax was introduced which is perfectly valid for methods like the above
add:def add(a, b) = a + bThe general rule here would be to only use one-line methods for methods that are functionally pure and avoid side effects. Conflating mutation and assignment in them would make those far more complicated:
def set_something(a) = something = a...like that's just confusing, and should likely be avoided in favor of using a traditional method.
There are two ways to specify a block function in Ruby, either with
{} or do .. end. As far as which one to use, that’s still quite the argument.10.times { |n| puts "The number is #{n}" }Typically it’s said that brace syntax should be used for one-liners
10.times do |n|
puts "The number is #{n}"
puts "Twice the number is #{n * 2}"
endand
do syntax should be used with multi-liners.Personally I have a habit of almost always using brace syntax for chaining reasons, but I’m also known to have some bad habits so take that with a grain of salt. The most similar style to the one I use is Jim Weirich's semantic blocks in which braces are used for blocks who's primary purpose is to return a value, and
dos are used for ones that have a primary purpose of side effects and output like so:[1,2,3].map { |v|
v * 2
}
[1, 2, 3].each do |v|
puts v
end...and Rubocop does have options for both, but the one-line versus multi-line is the default.
“More than anything else, code that looks like Ruby looks readable.” — Eloquent Ruby Ch 1
If anything, that quote should be framed above every Rubyist’s desk, but that’s a bit of a tangent.
We start with code like this:
doc.words.each do |word|
puts word
endwhich could be written instead like this:
doc.words.each { |word| puts word }Now, as mentioned earlier, puts is variadic, so it’d be more succinct to say it this way, but is less immediately clear what's going on:
puts *doc.wordsIt’s mentioned that when these expressions start getting a bit long, we should line break and go back to the
do notation:doc.words.each { |word| some_really_really_long_expression(with, lots, of, args).that(does_even_more(things)) }As far as how long is too long, the book mentions this is left to discretion. Typically it’s around 100 to 120 characters. Rubocop and the Ruby Style Guide reference 80 characters, but with modern 4k monitors and environments this can be a bit confining. To quote a certain pirate:
“And thirdly, the code is more what you’d call “guidelines” than actual rules.” — Barbossa
The book mentions two methods, which, individually make a lot of sense to omit parens from:
puts doc.author
doc.instance_of?(Document)but might not make much sense when chained together:
puts doc.instance_of? self.class.superclass.classNoted that recent versions of Ruby introduced
is_a? to be used as a shorter version of instance_of?.The book goes into looking at code in the wild for examples of how things like comments and examples are written. The example used is
Set in the book. A more recent set of libraries with good documentation I’d take a look into are ones like:Now there are others out there, and definitely take a gander at the Ruby Standard Library for examples, but this should give you some more ideas.
Notice that not all of them use YARDoc or other formatters. It comes down to a choice of exactly how detailed you want to be on documentation for your code. Rails doesn’t have very much in the way of inline documentation, but has a massive amount of documents written in markdown with examples everywhere.
Do what makes sense for your libraries.
Back to the chapter itself, they mention question mark and exclamation mark methods. Convention still follows that question mark methods return boolean responses, and exclamation marks typically indicate either mutation or surprising behavior.
There are some
Kernel methods like Float which break the rules, but are used as stand-ins for class names:# Works something like this:
module Kernel
def Float(v) = Float.new(v)
end
pi = Float('3.14159')There have been some cases in more modern Ruby of abusing bracket notation to avoid the necessity of a kernel method:
origin = Point[0, 0]which is defined like this:
class Point
def self.[](x, y)
new(x, y)
end
def initialize(x, y)
@x = x
@y = y
end
endThough this again is a matter of preference, as some would strongly prefer not to define methods on the top level namespace for their own gems and libraries.
“Above all else, pragmatism: You cannot make readable code by blindly following some rules.” — Eloquent Ruby Ch 1
Ruby is a language of flexibility and choice, following all the rules would quite simply be no fun and would lead to very rigid code. Do what makes sense and makes your code more readable, and if you’re confused feel free to reach out and ask.
Next up is control flow in Ruby!
34