My Thoughts on LangDev 2023

The LangDev 2023 conference, hosted by the Belastingdienst, in Utrecht, was all about web integration.

Another year has passed in the world of modelling and the grand wizards have convened, with some notable absences, to share their newfound knowledge.


LangDev 2023 was hosted by the Belastingdienst, or Dutch Tax Service, in their beautiful office in Utrecht. We saw a few Excel sheets in the spacious open offices, which felt more like a challenge to the DSL community than anything else.

The location was quite convenient for us, our boss had some unpaid taxes to take care of, and most of us were a short commute distance from it. That took away the magic of travelling, but I hear next one might be in Spain so we’ll catch up.


Day 1

The conference hit the ground running with presentations that were understandable, well executed and had compelling ideas behind them.

One thing I found interesting, as this is now the third language engineering conference I’ve been to, is that some of the presentations were like episodes in a series about the work of entire teams/companies. The Langium Talk by Irina Artemeva felt like that the most but also Wim Bast’s presentation on transformations.

I’d say the most frequent word of day 1 was "web". Everyone was rushing to move their sophisticated tools to the internet, so the users wouldn’t have to deal with complicated tools on their computers - they’d have to deal with complicated tools in a browser.

We got started off with an argument to use HTML form elements for projectional forms.

Then we had this "boring" presentation by Federico Tomassetti on parsers or something.

Then we saw how to Make models available on the web through Modelix which separates the users front end model editors and back end model storage and server, which creates the infrastructure for collaboration and lowers entry requirements for new users.

Followed up by what is in my opinion the best presentation: Dr Ambiguity by Jurgen Vinju who also had something to say on how "boring" Federico’s presentation really was.

And we ended the day with some more, you guessed it, web: An entire graphical language workbench, by Théo Giraudet, so I can teach my kids to do modelling without it looking boring to them, and Orca, by Pedro Molina, a graphical DSL for creating docker-compose configurations, so I can teach my kids how to use docker without it looking boring to them.

Moving to the web feels like a natural step forward, but todays internet has in many ways ruined my respect for browsers and their supposed proper functioning. Let’s see you adjust your web app 36 times for all the slightly different types of browsers, excluding the mobile versions. Although, further thinking about it, it would be funny to see how someone is frantically trying to fit in a language workbench on a smart phone.

User experience is set to improve in the coming years and it’s just going to cost a few front-end dev souls.

Day 2

Day two came around and moved us away from the web and shoved us straight into the nitty gritty details of some really cool projects.

Our host, Gert, presented how they built language translators which put me right back to school when I had to go and ask him to explain it like i’m five, because I couldn’t follow.

Our very own Niko Stotz, together with Jos Warmer, presented the LionWeb Initiative, a protocol that can connect modelling components, so people can work together and still use their favourite tools.

We saw that JetBrains is looking at moving MPS to the web and what the complexities would be when doing so.

We even got to see some wild code in its natural habitat, where the presenter, Wim Bast, with input from the audience, poked it and threw food at it to see what happens.

Being able to dive deeper into the presentations brought a sense of satisfaction. It felt like I did some leveling up in my understanding of modeling and development throughout the year, but there still were talks that I just sat there and participated in a staring contest with comprehension and lost. I’m not bothered by it, though, I’m looking forward to a time where I can also feel confident enough to ask smart people smart questions and have profound conversations about the answers.


I don’t know how many Model Driven/Language Engineering conferences were there before this one, and the following statement could be because of my own experience increasing, but this certainly felt like a more mature conference with practical projects and presentations that will be, or already are, used by an increasing amount of people.

There are calls to unite the community so it looks, and works, forward together, in a way that everyone does what they do best and everyone can use each other’s experience, and tools to get the job done.

I am looking forward to the next conference, not only because it could be in Spain, but because I want to see what new stuff people come up with and how the community grows.