Hi

Jfokus 2024 is done. I want to share some of my main takeaways with you.

Lots of focus on AI, both how we can use it and how it will affect our jobs. AI might not replace us now, but if we don't take advantage of it, we will be replaced by those who do. I will write more about this, so stay tuned.

I enjoyed listening to Chris Richardson talk about the trade-offs between monoliths and micro-services and how we can analyze different architectures to find what fits best for our specific case. He calls these things dark energy and dark matter. You can read more about it here: https://microservices.io/post/microservices/2021/11/30/dark-matter-dark-energy.html

Kevlin Henney had a workshop about how to write good unit tests. We often end up with tests that either don't test anything or don't reveal what they are testing. https://www.infoq.com/presentations/testing-communication/

Simon Brown talked about the software architecture role and that it's much more than choosing technology. It's about coding, coaching, and collaboration. Soft skills play a big part in this. A good architect needs soft skills such as leadership, communication, presentation, coaching, and mentoring. https://www.infoq.com/articles/architecture-five-things/

My presentation went well, and I was happy to have an almost full room. It's inspiring to see that more and more people are interested in growing their careers with the help of mentors. It was a bit of a challenge to condense the content into a 15-minute lightning talk, but it gave me a chance to practice being concise.

I also had the chance to meet and talk to many people. It is the best part of going to an in-person conference, so make sure you engage with new people when you go to one. You will learn so much and meet new people from all over the world. 

In my previous email, I wrote about the One Billion Row Challenge,  https://github.com/gunnarmorling/1brc, and at the speaker dinner, I had a chance to talk to Oleg Selajev from AtomicJar and the winner of that challenge, Thomas Würthinger from GraalVM. He had done some clever things to get the final execution time down from over 4 minutes to 1.5 s.

We discussed trade-offs, and he revealed that some decisions were made purely based on execution time. They did not know why it was the fastest. The resulting code is as effective as possible for the given problem, making it hard to change if the requirements change.

I talked with Charat Chandler from Oracle about community. It's important to constantly add new views and ideas to keep it alive and vibrant. I think that this is true regardless of the size of the community. If we don't evolve and grow, we will stagnate.

I talked about Swedish and Moroccan culture with Badr El Houari and had a chance to catch up with Venkat Subramaniam.

All in all, a successful conference. I learned lots, met old and new friends, got inspiration for new subjects to share, and hopefully shared something useful myself.

Enjoy

Markus Westergren