To put it another way, what does it take to learn a language? For a Spanish speaker learning Portuguese the process can be fairly straightforward: we need to understand the differences in pronunciation, understand the critical differences in vocabulary, and begin the more difficult process of understanding the differences in grammar. But what does the process look like for a French speaker trying to understand Sanskrit? Or cooking Biryani?
OK- so the last one is not a traditional linguistic language but I mention to emphasize the idea that sometimes learning a new language is not just a matter of picking up a new grammar with a new vocabulary. Sometimes we are approaching the issue of communication from a completely different perspective. I've seen many restaurant kitchens in my travels where the owner and the staff do not speak the same languages and yet they are able to convey recipes and tasks to another through the universal principles that we all can understand.
For the arrogant Java developer, looking to learn Scala, it's easy to imagine that the language can be learned by understanding the syntactical differences between the two languages. If our understanding of Scala is that it is a simple evolution of the Java language where we can avoid verbosity when declaring variables and utilize some syntactic-sugar to make our code more compact, then we might find ourselves quickly overwhelmed trying to learn the comprehensive features of Scala. If we simply scan through a summary blog post (wait, is that what this is?) with code examples, we might feel underwhelmed that the efforts in learning Scala won't yield any meaningful results. The arrogant Developer has years of experience and wants to learn the language now!. Why does it have to be so difficult?
The answer lies in the intention of the Scala language. As its creator, Martin Odersky, put it in an interview:
Scala was designed to show that a fusion of functional and object-oriented programming is possible and practical." If we understand this to be the motivation we can set the proper intention for our study of Scala: when a developer learns Scala they are not just learning a new language, but an approach to developing algorithms, data modeling, and design with the most concise and efficient object-oriented and Functional design.
The arrogant developer might have years of OO design under their belt, they may have memorized books of Design Patterns and might even know how to write some functional expressions with Java 8's new Lambda and Streams features. But Scala takes that to the next level with language features that support Algebraic Data Types, implicit parameters, monads, and more. As someone who is going down the beautiful rabbit hole of learning a new way of marrying functional with object-oriented programming, I wanted to put together a syllabus for experienced developers who want to learn a new language.
This blog post is not a tutorial on how to learn Scala, that is a big task, but rather a list of concepts to study to understand not just how to program in scala, but why to do so in a specific manner. When approached this way you will learn many new concepts and different ways of writing code, managing side effects and more. The arrogant developer will probably find themselves being quickly humbled at the breadth of it all,
a testament to the language and its community of developers.
Let's get started!
Adopt the right mindset
(Re)learn functional programming
Apply these differences to your object-oriented principles
Alway Be Coding
Humility, creativity, and discipline. Check your ego at the door and prepare to realize that many of the concepts that make for a good scala developer could be unfamiliar to you.
There have been many developments over the years in the world of functional programming and if you have not been keeping up then many of this will be new ground.
Scala is not "better Java": This is not Kotlin which attempts to erase some of the verbosity of the Java language and provide language features that facilitate better programming... although that is part of it. The primary goal of Scala is to facilitate Functional and Object-Oriented principles in a JVM based language.
The Functional Programming Landscape is Complex: There are many concepts that will most likely be unfamiliar to you. If you find yourself gravitating towards a resource like Scala for the Impatient you might be skipping out on learning some of the functional paradigms that will help you on the larger journey of becoming a better programmer and taking full advantage of everything Scala has to offer.
Functional programming has been described many different ways. I have always considered it to be the creation and composition of rules to describe a system. In this context, rules are functions: f(x) = y for example, tells us that when we supply input x into the function we get y every time. From a mathematics perspective that is great. But the astute computer scientist and the practical computer engineer will observe that things are seldom so simple. If our function has to fetch data from a database, how do we react to deadlocks? Can these rules work in a multi-threaded environment with mutable objects? Deciding how to manage the theory of functional programming with practical realities is not easy and Scala takes specific steps to address that.
Good functional programming skills can have a profound influence on your OO development while the converse is not necessarily true.
Once you understand the concept of pure functions and what they are, you will know a great deal of how Scala manages its functional state. At then end of it, you should know:
What pure functions are and how to create them.
What mutation/mutability is, what problems they can cause, and how pure functions avoid these problems
What are side-effects in a function? How does Scala deal with them?
Section 2.3, Pure Functions for Domain Behavior from Functional and Reactive Domain Modeling
Scala Best Practices: Pure Functions at DZone.com
A side effect of worrying about side-effects can be sweaty palms and headaches. Scala tries to mitigate these symptoms with built-in operators that help manage state when things go wrong so don't get discouraged and start looking at how you can streamline your own functions and methods to bring them closer to purity.
How smart will you appear at your next job interview if you start talking about algebraic data types and no one else knows what you are talking about? I mean, the word "algebra" is in there. Algebra! Purity of functions is great, but now we move into purity of our data types. When you learn about ADT you should know:
The two types of ADT: sum and product, and how each one is instantiated.
How Scala's traits and case classes relate to sums and products.
Why immutability is important with ADTs and how Scala's ADT design favors immutability.
Chapter 4, Modelling Data With Traits from Essential Scala
Algebraic Data Types in Scala from Alvin Alexander's Blog
What other languages on the JVM provide first-class support for ADTs?
Rare is the driver that will go through life without a speeding ticket. Similarly, no Java developer will go through life without running into a few checked exceptions here and there. Both are equally annoying and exhausting. In addition to wearing out keyboards, the verbosity in writing out try/catch/finally logic does not play well with the functional programming paradigms you were learning about in the previous sections.
Scala provides native types to facilitate things here that act as wrappers around data similar to the way Java's Optional works. After studying Scala's wrapper types you should be familiar with:
Try and how it can be used in situations where exceptions can occur.
Optional and how it's implemented similarly as in Java (but integrated more
tightly into the Scala API).
Future and how it can be used with potentially higher-latency
operations we want to manage asynchronously.
Either, the red-headed step-child of Try and the problems that can occur in trying to get this to compose (but why you still may need to use it).
Section2.6 Making models reactive with Scala from Functional and Reactive Domain Modeling
Everything you learned studying pure functions and ADTs leads directly into the concept of monads. You will know you have a proper understanding of monads when you are familiar with the three components of a monad:
How it's defined using a trait or case class.
How it's functionality is coded in its constructor for parameter x and it returns its possibility of y.
The relevance of flatMap for accessing y and utilizing this monad as part of an aggregate.
Section 6.6 Monads from Essential Scala
Try reading What the heck is a monad?, because Monads aren't too difficult to understand, but can be kind of a nightmare to explain! Scala's implementation of monads can differ slightly from other implementations (Haskell, for example) adding to the confusion.
For most Java developers, making your way through learning about the Scala approach to Functional programming will be the toughest part of the curriculum. At that point you can focus on some of the OO features Scala offers which feel like welcome extensions to the language. Managing side effects makes your code cleaner, more concise, and easier to test. In addition, Scala offers several language features that make implementing certain patterns easier and more readable than Java. A Kotlin developer will probably recognize some of those concepts and Java itself is also trying to incorporate some of these (albeit slowly).
A very simple concept that has a very commonplace implementation. You should know how it works and how it relates to the Factory pattern in Java.
the unapply method for deriving an object from a parameter.
While Scala tends to encourage immutability with some of its constructs and conventions, developers that are not yet ready to give up the ole setter method can still follow a traditional approach, while the native copy method provides a simple construct for creating cloned modified instances of existing objects.
How case classes come with built-in support for copy and how updates can be communicated via copy.
How copy works for complex objects with deep nesting. Also, how lenses can be used for trickier updates.
How to use var to permit mutability of a field, and what are some of the valid use cases to consider this?
Java 9's modules are the most notable feature of this release and can find some symmetry in how scala allows for fine-grained control over access to a class' data.
How Scala can restrict access to methods using private and protected and how protected differs from Java to Scala.
How private package works and can be used to restrict access to a different package than the current class.
How we manage Lists, Maps, Sets, and other data structures is always a huge part of any language and Scala is no different. Because it supports integration with Java, Scala's collections implementation is fairly similar to Java's implementation, but contains many enhancements that make working with it quite nice.
How to distinguish mutable collections from immutable ones.
One liners to declare, map, and filter a given collection.
Using zip to combine two collections.
I feel that strongly understanding these Object Oriented features can be derived from going through Essential Scala and working the examples therein.
It's always important to work those keyboards and put the principles you are learning into practice. Configure your IDE to use Scala and get started! One of the great things about Scala is that it has worksheets which act as a command shell for instant evaluation. This allows you to test out new concepts quickly.
99 Scala Problems is a great collection of problems that will test your understanding of Scala's syntax against some challenging algorithmic exercises.
I found that Essential Scala had the most natural flow of exercises for learning, although they could get tedious at times.
This online PDF provides an excellent overview of the Scala language and its features. It does not necessarily emphasize the functional model of doing things but gives you all the tools you need to be a great Scala developer in all disciplines.
If this book was referenced ubiquitously throughout this post, it is because it is a great resource for all developers. It's a Scala book disguised as a bible on Functional programming and domain modeling. The result is you can learn a lot of best practices in these fields while learning how Scala correlates to them.
Written by the creator of Scala, Martin Odersky, this text is a comprehensive guide to Scala. I don't think it is organized as effectively for Java developers from the old and new schools, but you would not be amiss to have it on your shelf.
Taught by the creator of the language and his team, this Coursera class is a detailed deep-dive into the language and is not for beginners. In addition to Functional Programming concepts, it touches on using Scala for parallelism and big data analysis.