Report: Blue Harvest at JSConf Belgium 2019
Last month I had the chance to go to the JS Conf Belgium 2019. The conference was in the city of Brugge, which in total was about 3 hours travel by train from Amsterdam. The venue was located just in front of the station at Howest RSS 1, one of the buildings of the Hogeschool Howest. The location was big but only two rooms were set up for the talks, placed on different floors separated by narrow corridors. There were no sponsors or business stands in the venue, so it was essentially the talks.
This was more an annoyance (I enjoy visiting the business stands during breaks) than an issue. The talks were good and varied, and only a small minority were a blatant marketing stunt.
Because of my train, I missed the first talk, To all the code I’ve loved before by Yuraima Estevez, so I started with the choice between Scaling JavaScript to Video Ad Tech by Tim de Pauw and 2nd Generation JavaScript Frameworks & Libraries by Geertjan Wielenga. I chose the later for the next presentation, and although the real purpose of the talk was to promote Oracle’s front end framework JET, I really enjoyed the insights from Oracle, one of the biggest corporate providers, about what corporate partners look for when choosing certain tools or frameworks. We will make good use of those insights to shape our strategy at Blue Harvest so stay tuned for upcoming exciting news in the near future!
The second choice was between Let’s compose Web components by Henrique Ramos Limas or From 120 to 800 MB of RAM usage in 5 minutes. A story of plugging memory leaks from Rafal Rumanek. Given that I recently had the chance to attend to the talk about web components from my colleagues Carel Coenraad and Gustavo Bremm, I chose to learn more about memory leaks. Rafael taught us about how much 100ms user interactions matter and to be aware of event listeners, destroyers and long tasks. Valuable information when designing front end applications.
The next choice was a tough one: Back to the future of JS II: Beyond what we can foresee by Willian Martins da Silva or JavaScript WebRTC: make JS Audio & Video Phone calls from your Web App by Alessandro Polidori. Both looked super interesting! At the end, my choice was with the first talk, given my personal and academic interest in the future of JavaScript as a language. That future seems quite bright with proposals like pattern matching, which something very common in Haskell and Rust, pipeline operator (all in into the UNIX way!), or binary AST, which will allow us to have fast loading times for front end applications or fast cold boot ups if you are into server-less computing.
There was no hard choices after, being the only option to assist to Tom van Cutsem’s Browsing NPM packages more effectively with Code Compass. This talk was sadly a promotional talk from Nokia to let people know about their tool Code Compass for discovering NPM packages. Not really interesting.
After that unfortunate talk, and a refreshing tasty lunch, the next talks were TDCD: Test Driven Component Development., by Tim Doherty, and Passwords are so 1990 by Sam Bellen. Given my interest in cryptography, I chose to assist to Sam Bellen’s talk. It was the right choice: in a very amicable presentation, he showed the downsides of the previous identity methods like cookies in a modern web full of Single Page Applications (SPAs) that rely on having Cross Origin Resource Sharing (CORS) capabilities, to follow with a modern approach using JSON Web Tokens (JWT) and OAuth, to wrap up showing what the future might be with WebAuthn.
The next choice was between Machine learning for web developers by Lander Verhack, and Taking your Angular application to the next level by Martijn Werbrouck and Matthias Goderis. Given my interest in artificial intelligence and that I am not an Angular developer, I chose the first option. The talk was very light but deep at the same time. It started clarifying what is and what is not machine learning, and a brief explanation about which are the main concepts behind it; to continue about what is needed to perform machine learning and how we can do it from a browser relying on Tensorflow.js and WebGL instead of the more classical combination of Python with CUDA.
I am still disappointed about my next choice: The search for App-iness : Progressive Web Apps by Jamie Maria Schouren versus Transforming your hand drawings into HTML with AI by Nick Throgh. I have built previously some PWA in the past, so I did not expect to get as much new content as from the second talk. Sadly, I ended up in a pure marketing talk about Azure services from Microsoft. A waste of time.
After the deception of the previous talk, I got to choose between Making React applications work for everyone by Yuraima Estevez and Contributing your time and effort to Open Source Software by Diedrik de Mits. I found more interesting to learn about accessibility (a11y), given that I have already contributed to some open source projects. It opened my eyes about how other people perceive the web and how it is important to make sure our web content is accessible for everyone. Although it can be daunting at the beginning, it can be summarized in 5 easy rules:
- Provide a text alternative to all images and videos.
- Make the website operative with only the keyboard.
- All the form elements must have a label tag.
- Make the website robust and understandable using ARIA attributes.
- Write valid code.
The next session I had the choice of learning more about Node.js DevOps at Automating your NodeJS software delivery by Lieven Doclo, or more about components at Don’t make component abstract art by Haroen Viaene. I have a preference for back end development rather than front end development, and my current project is about DevOps, but the presentation looked too much like a marketing talk, so I chose the second one. The presentation was not the greatest, but it presented a way about how to make good abstractions in components that makes them easy to add, easy to override and prevent mistakes. Sadly, the indications about how to achieve these abstraction were also a bit abstract.
Finally, the last choice was between Building dumb sh*t from Sara Vieira and JavaScript on the Internet of Things by Sille van Landschoot. Lured by the title, I chose the first one. It was not a technical talk, but it was a very inspiring one. Building random stuff that has no use is great for several reasons. It will not make us rich, but we will learn something new, or at least, laugh a bit. That is why you should go now and do build some random stuff.
In summary, the conference was good, with some improvement points. Talks were interesting but not deep. Not too many blatant marketing talks, although some of them were unavoidable. The place was well located but not well prepared, and there was little to do in between talks. The conference was fairly priced and offered possibilities of discounts for underrepresented collectives. Definitely worth to check for the next edition.
(All the pictures are part of the official photos taken during the event, and they belong to jsconf.be Thank you very much JSConf Belgium for allowing us to use them!)
