Ruby is one of the most expressive languages that I ever worked with. What’s funnier about that is the paradox: it can be as beaultiful or as cryptic as you want. It’s not a foolproof language either, nor does it try to protect the developer from his own mistakes - it’s a sharp tool!
Ruby is a multi-paradigm programming language; it lets you write code in many different styles: procedural, object oriented, functional, metaprogramming and others. You can write cryptic Ruby code if you want, as it still has some Perl and other obscure inheritances, but I prefer to avoid that style whenever possible. The Ruby way says we should have more than one way to write the same code, and that’s a good thing! Unless you are writing a quick command line hack, you’ll most likely want your code to be clear and expressive.
Mostly everything in Ruby is a method call, be it implicit or explicit, except for some few language constructs. You may know that Ruby allows you to make method calls without parenthesis. What’s the point of that? Expressiveness, of course!
First of all, I read the line out loud: does it better express the intent with or without parenthesis?
With the parenthesis
Take that piece of code:
Read the body out loud: it reads like plain english, and the lack of parenthesis reinforces that feeling. At first glance it seems like it’s telling “hey @names, update the additional names!”… so, will our names add something to the additional_names collection? It doesn’t make sense! Of course we know what this line is supposed to do, but the real intent would be clearer at first glance if we could write it like this:
Now let’s add the parenthesis:
The difference here is that the parenthesis makes it clear that we are working with an expression, a mechanical method call that is not supposed to read like english, reinforcing the idea that additional_names are being added to names, and not the other way around. The bottom line is that the english sentence gets confusing without the parenthesis.
The update method is part of the Ruby Hash class, and it’s an alias for the merge! method. You will find that kind of thing in Ruby a lot. The language makes it very easy to alias methods, and encourages you to do so. There are many methods with different names which do the same thing. For example, take map and collect: both methods iterate through an array, transform its values into something else, and return a new array at the end. The point of having both is that there are situations where collect is more readable, and others where map wins. Intent is the word. We could go on with a big list: inject and reduce, detect and find, etc.
Another situation where I prefer parenthesis is when the return value is meaningful, because it’s clear something will be returned to the caller:
Now let’s take a look at another example:
There is also a return value here, so it reads really well. Even if we tried to run it without the parenthesis we would get a warning, because the & character would confuse the Ruby parser. Ruby tries to make an assumption about what you want, but it isn’t entirely sure whether you meant to use an overloaded operator or the Symbol#to_proc syntax. It tries to be useful, pointing out for potential errors in your code. Note that we need to run the Ruby interpreter with the -w flag to see that warning.
Without the parenthesis
Method calls written without the parenthesis are essentially imperative calls where the return value doesn’t matter. For example:
Of course, there are some situations where we can’t do that, and for a good reason. Ruby emits a warning if you try to use the following syntax:
For the same reason described above, that line confuses the Ruby parser: who knows, you could be trying to perform a division between two variables. So we are obliged to include a parenthesis in that method call:
Conclusion
Everybody has its own guidelines when it comes to expressiveness, but the style presented here aims to be consistent. Some people simply hate parenthesis, taking them out whenever they can - even in method definitions, which I personally don’t like! Parenthesis in method definitions add structure and make the code easier to read, effectivelly communicating “this is a method body”. Let’s not be too radical, OK? Parenthesis actually have a purpose!