First Julia workshop at ARC

Last Friday (the 10th of November) we run our first ever Julia workshop. After years of having an expert in our team – Mosè – who has been introducing the rest of the team to this wonderful language and even convinced some collaborators to use it on their projects, we’ve done the jump to teach it to the UCL community.

This first workshop was limited to a reduced number of learners (10-12). Seven of which attended (most of them physicists!), with our team of three (myself instructing, with Mosè and Tuomas helping — also all physicists 😅) made the learners’ experience very positive.

Originally, we were going to use the Carpentries Julia lesson available in the incubator. However, Mosè and I decided against it as the expected previous knowledge was higher than what we were aiming for. Therefore, we created our own lesson!

Our lesson started with the basics, different types of numbers, strings and how all them fit in the family of types in Julia. We introduced some of the quirks Julia surprises you with when you come from a different language. This was key in our lesson! We started to write a function as if it was Python — which was what we expected to be the most familiar for our cohort. From there, we were introducing new concepts and syntax to make our code more “julianic” (I’ve come up with that term, so it may not be the one used by the Julia community). We covered the basics (types, function, conditionals, loops and plotting) during the morning session. After lunch, we went to introduce how to use other libraries to solve polynomials and ordinary differential equations. We even introduced unit testing and had time to learn how to work with CSV files with DataFrames and gave a quick overview of Pluto.

During the preparation of the material and the class, I was constantly supported by Mosè, bouncing lots of ideas and suggestions. We’ve even found a bug in one of the libraries we were going to use that they fixed instantly after Mosè reported it.

The class went smoothly. We encountered some problems with the installation of Julia and some unexpected slowness when installing libraries (we reported it after the workshop, and it was also fixed straight away!). This is some of the feedback we’ve received at the end of the day:

• Great course, learned a lot.
• The course has been great. The pace is good and it allows us to ask any questions we have.
• Comparison’s to Python really helped me appreciate the advantages of Julia. Paper plane example was great.
• Very good course, covered all the right topics for a 1-day intro session.

Personally, I don’t remember a class that has gone so well! With very little difficulties, covering everything we were planning to do and answering very interesting questions from our learners. It may have been due to the small number of learners, or because of their previous programming experience, or the similar background across all of them, or maybe, it’s because Julia is easy to learn 😉. Whatever reason it is, I really want to repeat it, with a larger class and a more varied background of learners. There’s no reason for only letting the physicists have fun with Julia, right.

So, if you are interested in learning Julia, be sure that we will repeat more sessions like this one! This may be too basic for you? Don’t worry, we are also planning to run a more advanced workshop focused on Julia for HPC during Term 2’s reading week. Keep an eye out for our future announcements.

Now that we have started, we won’t stop!