Preface

Julia is a relatively young programming language. The initial design work on the Julia project began at MIT in August 2009, and by February 2012, it became open source. It is largely the work of three developers Stefan Karpinski, Jeff Bezanson, and Viral Shah. These three, together with Alan Edelman, still remain actively committed to Julia and MIT currently hosts a variety of courses in Julia, many of which are available over the Internet.

Initially, Julia was envisaged by the designers as a scientific language sufficiently rapid to make the necessity of modeling in an interactive language and subsequently having to redevelop in a compiled language, such as C or Fortran. At that time the major scientific languages were propriety ones such as MATLAB and Mathematica, and were relatively slow. There were clones of these languages in the open source domain, such as GNU Octave and Scilab, but these were even slower. When it launched, the community saw Julia as a replacement for MATLAB, but this is not exactly case. Although the syntax of Julia is similar to MATLAB, so much so that anyone competent in MATLAB can easily learn Julia, it was not designed as a clone. It is a more feature-rich language with many significant differences that will be discussed in depth later.

The period since 2009 has seen the rise of two new computing disciplines: big data/cloud computing, and data science. Big data processing on Hadoop is conventionally seen as the realm of Java programming, since Hadoop runs on the Java virtual machine. It is, of course, possible to process big data by using programming languages other than those that are Java-based and utilize the streaming-jar paradigm and Julia can be used in a way similar to C++, C#, and Python.

The emergence of data science heralded the use of programming languages that were simple for analysts with some programming skills but who were not principally programmers. The two languages that stepped up to fill the breach have been R and Python. Both of these are relatively old with their origins back in the 1990s. However, the popularity of these two has seen a rapid growth, ironically from around the time when Julia was introduced to the world. Even so, with such estimated and staid opposition, Julia has excited the scientific programming community and continues to make inroads in this space.

The aim of this book is to cover all aspects of Julia that make it appealing to the data scientist. The language is evolving quickly. Binary distributions are available for Linux, Mac OS X, and Linux, but these will lag behind the current sources. So, to do some serious work with Julia, it is important to understand how to obtain and build a running system from source. In addition, there are interactive development environments available for Julia and the book will discuss both the Jupyter and Juno IDEs.