One foundation of Functional Programming is Referential Transparency (RT). Purely functional languages (like Haskell), and purely functional libraries for non purely functional languages (like scalaz or cats for Scala) aim at building Referential Transparent programs. The benefit of this constraint might not be immediately visible, so in this post I want to expose my thoughts on this, and see what Referential Transparency can buy us.


Any application in an enterprise context, regardless how small this context might be, must relate with other systems. These systems might be File system, databases, webservices, message queues, logging systems, or systems using a particular communication protocol. Moreover, data typically undergo transformations, switching and routing logics before reaching other systems. The variety of combinations this allows is enormous, and tackling each of these in a hand made, custom way might easily become an integration nightmare. Enterprise Integration Patterns (EIP) establish a standard way to describe and identify the different approaches that one can follow to deal with an integration problem (see They establish a common vocabulary that can be used unambiguously when talking about integration. If we consider that integration solutions are ubuquitous in application development, we realize easily how convenient it might be to have solid foundations on this subject.

After the Scaladays 2015 in SF, a lot of echo has involved the Scala.js framework. Scala.js compiles Scala code into Javascript code, which can be run in a browser but also on Javascript powered server environments e.g. Node.js.

I embraced functional programming through Scala starting with the great Marting Odersky’s course on Coursera, following with the Reactive Programming course in Coursera (a second edition of which will start shortly and I really recommend you to sign up and follow it), and later on working with Play framework and Akka streams.

Why learning concurrency

Studying for the OCP exam offers a nice opportunity to dive into one of the most important (and sometimes overlooked) aspects of the Java language and JVM: threads and concurrency frameworks. Before I never delved into the details about concurrency just because I have been working on a container (like Tomcat) or on a framework (like Apache Camel) that takes care of distributing load on a thread pool, and basically every request or every message processing can in 99% of cases be treated as a synchronous process, so I survived for many years without having a good understanding of how concurrency works in Java.