Title: A Glint of Ruby

Author: Bob Schmidt

Date: 06 September 2017 16:23:46 +01:00 or Wed, 06 September 2017 16:23:46 +01:00

Summary: Pete Cordell shares his experiences with learning a new scripting language.

Body:

I mainly program in C++. I also use C# for Windows GUI programming, PHP for web work and a scripting language for odd tasks such as analysing text files or running tests.

I have been using Perl in the scripting language role. However, even I had to admit that that was getting long in the tooth and it would be good for my skill set to move onto something more modern. The logical choice was Python. Iâ€™ll admit that I never really got on with Python. The sorts of things I wanted to do often seemed harder to do in Python than in Perl. But Python seemed to be the way things should be done, so I decided to persevere. Then, by chance, I got involved in a project with someone who was using Ruby [1]. After an initial â€˜whateverâ€™, it has been love at first sight ever since!

I tell you this to put the rest of the article into context. Iâ€™m not an experienced Ruby programmer, and like an infatuated teenager, I may well be blind to any warts the language might have.

There are many Ruby tutorials online, so I donâ€™t intend to teach you Ruby here. Instead, my aim is to help you recognise a Ruby program when you see one, try to pique your interest in learning more about Ruby, and cover a few things that might be intriguing to a C++ (and possibly Python) programmer not familiar with Ruby.

You may already have Ruby on your system if you are using Linux. If not, you should find a suitable download option at [2]. In addition to providing the ability to run Ruby programs on your system, you should also find an interactive Ruby shell called irb. This is very handy for experimenting with Ruby, and you may wish to use it to play around with some of the examples below (most of which you should find at [3]).

A contrived example

The obligatory â€œHello, World!â€ program looks as follows in Ruby:

  puts "Hello, World!"

Itâ€™s always a good sign when a scripting language allows â€œHello, World!â€ to be a one-liner, but it doesnâ€™t tell us much about the language. So, in the spirit of feature creep, letâ€™s look at Listing 1.

# A class to say Hello
class Greeter
  def initialize( who )
    @who = who
  end
  def greet
    if @who =~ /^\s*PETE\s*$/i
      puts "Hello author"
    else
      named_greeting
    end
  end
  private def named_greeting
    puts "Hello #{@who}"
  end
end
greeter = Greeter.new "reader"
greeter.greet
Greeter.new( " Pete " ).greet

Listing 1

Working through the example, comments start with the # character and continue to the end of the line.

Class names (i.e. Greeter) must begin with a capital letter, and, by convention, use PascalCase.

Chunks of code (a â€˜blockâ€™ has a special meaning in Ruby so Iâ€™m avoiding its usage here) typically start with a keyword such as class, def, if, while and continue until a matching end keyword.

Indenting is not particularly important in Ruby. The Ruby style guide specifies two spaces per tab. Iâ€™m yet to get comfortable with that, but Iâ€™ve adopted it here as itâ€™s more idiomatic, and better for presentation in a print publication.

The def keyword defines a method. Method names typically are all lower-case and sub-words are underscore separated. The last character can also be one of ?, ! or =. The initialize method is analogous to a constructor in C++, and is called when a new instance of an object is created.

The scope of variables in Ruby is indicated by an optional prefixed character. Without a prefix, a variable is local to the chunk it is defined in. Variables prefixed by a single @ character have object instance scope. Variables prefixed by @@ have class scope, similar to C++ static class variables. Variables prefixed by $ are global. So, the line @who = who translated to Python would look like self.who = who. Variables names must begin with a lower-case letter or underscore, and, by convention, are all lower-case with sub-words separated by underscores (e.g. my_variable). (Constants on the other hand must start with an upper-case letter, and are usually all upper-case along with underscore separators, e.g. MY_CONSTANT.)

Moving onto the greet method, we get our first glimpse of how minimalist and token free Ruby can be. There are no brackets around the if clause and there is no delimiter such as then or : to explicitly mark the end of the condition. One of the goals of the Ruby designer (Yukihiro Matsumoto, or Matz to his friends) was to make programming faster and easier. One aspect of this is to make the interpreter work harder in order to make life for the programmer easier. Typically a Ruby expression ends at the end of a line, unless there is some kind of binary operator, or a , to suggest that there is more on the next line. Sometimes additional brackets are needed for disambiguation (and semi-colons can be used to terminate expressions), but the preferred Ruby style is to avoid them if possible. This can take a bit of getting used to, but after a while the result looks more narrative like than typical C++ programs.

The condition of the if expression shows that regular expressions have first-class support in Ruby. Those with a Perl background will find this syntax familiar.

The puts method call should be familiar to C++ programmers, although you may be excused for having forgotten about it. puts prints its arguments to standard output, followed by a newline sequence. If you donâ€™t want the newline characters, use print instead.

Moving to the else clause we have a call to the named_greeting method. Unlike in Python, there is no need to prefix a call to another method in the same object with self..

The named_greeting method definition itself is preceded by the private method, showing Ruby gives you the ability to control access to object methods. Note that private isnâ€™t a keyword like in C++, but a method call. Thereâ€™s some Ruby subtlety here which is too detailed to go into in this article. Suffice to say that the use of private on the same line as a def will make only that method private (similar to Java and C#â€™s usage), whereas private on its own line (a call to private with no arguments) will make all following methods private (similar to C++â€™s usage).

The puts line in named_greeting shows Rubyâ€™s syntax for string interpolation. Many scripting languages support string interpolation, but Ruby takes this to the max. The code between the opening #{ and closing } can be any expression and is not limited to just variables. It can include operators, method calls, and even if statements. If the result of the expression is not a string, then the interpreter will automatically convert it to one.

The class definition described creates a runtime object called Greeter. That object includes a method called new. When the new method is called on the Greeter object, it creates a new object that conforms to the class definition. It then calls the newly created objectâ€™s initialize method with the arguments given in the new method call. Note again that there are no brackets around the arguments to the new method call.

In the example, we assign a reference to the new object to the greeter variable. Once we have a reference to an object instance, we can invoke methods on it as shown by the expression greeter.greet.

As you might expect, Ruby is garbage collected so thereâ€™s no need to clean up any objects that have been created.

Hopefully it wonâ€™t be a surprise to you that the output of the program is:

Hello reader

Hello author

Digging for gems

Now you hopefully have a feel for what a Ruby program looks like, letâ€™s have a closer look at a few things that might be intriguing to a programmer who only knows C++, and things that are more unique to Ruby.

Objects everywhere

Everything in Ruby is an object. There are no primitive, machine word level types such as char, int and float like you find in C++. Consequently, you can call methods directly on explicit values such as floating-point numbers. For example, 1.5.round evaluates to 2.

All classes are open to extension in Ruby and are similar to C# partial classes. For example, you can augment the Float class with your own methods (although this isnâ€™t a recommendation to do it). The following program outputs 4.5:

  class Float
    def triple
      3 * self
    end
  end
  puts 1.5.triple

It may look like the usage of puts above is a free-standing function. But it is actually a member of the Object class from which all objects derive. The other â€˜fudgeâ€™ is that the outer-most level of execution takes place within the scope of an automatically defined, but largely hidden, object called main.

Types and whatâ€™s true

In addition to numbers, strings and the regular expressions already mentioned, Ruby also supports arrays and hashes using the handy syntax youâ€™d expect from a scripting language. Additionally, Ruby has a range type which captures a start value and an end value within a single object. Iâ€™ll touch on ranges in a bit, but hereâ€™s some examples of the syntax for these types:

  my_array = [ 1, 2, 3, "Four", 5 ]
  my_hash = { "author" => "Pete",
              "reader" => "you" }
  my_range = 0..10

Ruby will automatically convert between numbers of different types (for example, integer to floating point number) as the operations require. But unlike some languages, outside of string interpolation, it wonâ€™t automatically convert to and from numerical types and strings.

As with Python, but not with Perl, Ruby has a Bignum type that can store integers of any size. If integer operations get too big to fit within a native machine sized integer, they will be transparently converted to a Bignum. This is shown in the following code, where ** is the â€˜to the power ofâ€™ operator, and the output is shown in the comments:

  puts 2**4          # 16
  puts (2**4).class  # Fixnum
  puts 2**70         # 1180591620717411303424
  puts (2**70).class # Bignum
  puts (2**70/
        2**60).class # Fixnum

Then there is true, false and nil. nil is equivalent to C#â€™s null and Pythonâ€™s None. Only false and nil are treated as false in conditional expressions. Everything else is treated as true, including 0, empty strings and empty arrays. Those familiar with Python and Perl might find the latter surprising, but it does seem to fit in with the language quite well. If nothing else, itâ€™s easy to remember!

The block

One of the more unique features of Ruby is the concept of a â€˜blockâ€™. Similar to lambdas, any method can be given a block which it can call with a set of parameters, and receive a result back. The â€˜blockâ€™ follows a method call, and is either contained between opening and closing braces, or the do and end keywords. The former syntax is typically used for one-line (or even in-line) blocks, and the latter for multi-line blocks. The arguments to the block are delimited by a pair of | pipe characters, and the value of the last expression executed in the block is returned to the calling method. For example, using the variables set up previously:

  my_array.each { |x| print "_#{x}" }
  puts  # Put new line at end of above
  my_hash.each do |role, name|
    if role == "reader"
      puts "#{name}, Thanks for reading"
    end
  end

Outputs:

  _1_2_3_Four_5
  you, Thanks for reading

Blocks are not limited to iterating through arrays and hashes.

If the File.open method is given a block, it will close the opened file on return from the block. As such, this usage scenario is similar to C#â€™s using and Pythonâ€™s with constructs.

In combination with suitable methods, blocks are often used as an alternative to conventional counted for loops. In place of C++ like:

  for (int i=0; i<10; ++i)
    std::cout << "_" << i;

you can do any of these (wherein the last one is an example of the range type I mentioned earlier):

  10.times { |x| print "_#{x}" }
  0.upto(9) { |x| print "_#{x}" }
  (0..9).each { |x| print "_#{x}" }

Blocks, and the fact that most methods return something â€˜usefulâ€™, opens up the way to a more functional coding style. Among many functional-like methods, Ruby supports map, select and reduce on arrays. Using these, if you wanted to know the sum of the odd squared numbers, for the numbers 0 to 100, you could do:

  puts (0..100).map { |x| x * x }
    .select { |x| x.odd? }
    .reduce { |acc,x| acc + x }

Why Ruby uses select rather than the more traditional filter, Iâ€™m not sure. If it bothers you, though, you can do class Array; alias filter select; end and use filter in your code instead.

One thing that perhaps should be said here is that I havenâ€™t been able to see how to make an expression like the above â€˜lazyâ€™ in the same way that Haskell does. Each method call creates a new array before the next method acts on it, rather than just acting on iterators. This wonâ€™t cause problems in many cases, but does prevent Haskell-like solutions of the form â€˜for all positive integersâ€¦â€™.

Thereâ€™s moreâ€¦

Thereâ€™s much more to Ruby than Iâ€™ve been able to cover here. Ruby has parallel assignment, slices, exceptions, threads and coroutines.

It also has an extensive library of third-party code in the form of RubyGems [4], and a dependency management system called Bundler [5] that ensures that the correct versions of the Gems are installed for your program.

Finding out more

If you want to explore further, there are numerous Ruby tutorials on-line (e.g. [6]). I have found The Ruby Programming Language book [7] to be very good. This is a depth-first look at the language that assumes you already know a bit about programming. It feels as close to a language definition as you can get without having to be a language lawyer. The Ruby documentation [8] is good, but I find it hard to navigate. Hence, I tend to use Google as itâ€™s front-end! Googling will also bring you to Stack Overflow. Perhaps because Ruby is slightly less common than other languages, the answers for Ruby seem to be of a higher quality than you might find for other languages. The irb interactive Ruby shell is also a fun way to poke around to test your understanding.

Conclusion

I hope Iâ€™ve given you enough to be able to recognise a Ruby program. Possibly youâ€™re now interested in finding out a little more information. I can understand that some people might find the various naming constraints difficult to get over, but in my case I had already adopted similar rules, and so wasnâ€™t bothered. My biggest gripe is that if your code changes a constant, the Ruby interpreter will only generate a warning rather than an error. I console myself that, for the sorts of things I use Ruby for, I donâ€™t need a lot of constants. Those wanting to make more extensive use of Ruby might consider this more of an issue.

Overall, Iâ€™ve found code solving similar problems is simpler and clearer in Ruby than in other languages. As such, it has exceeded my expectations and I hope it will for you too.