Blogs, terminology, amazement

I always look forward to the monthly JuliaComputing blogs, and I tend to point people to them (and e.g. the amazing case studies). They are usually readable.

In the latest September one there is mention of:

NeuralFMU: Towards Structural Integration of FMUs into Neural Networks - Tobias Thummerer, Lars Mikelsons and Josef Kircher

FMU is not explained in the title or abstract (nor JuliaComputing blog):

Keywords NeuralFMU, FMI, FMU, Julia, NeuralODE

I might be speaking for many people when Julia is the only recognizable term, if that. NeuralODE was new to me/all of us(?) not long ago.

It made me recall from Matt @mbauman we are all in a 1% exclusive club (those who just know how to program anything), from his amazing early (JuliaCon?) video. I wander how exclusive the more specific science/Julia language club is, or e.g. knowing about NeuralFMUs.

What I most like about Julia is that I’m constantly learning of new stuff, new areas of science and math. I can’t imagine any other programming language community/ecosystem having as much coverage and depth.

Not too long ago (or before Julia), I did not know what e.g. automatic differentiation was, and I admit reading about Julia could be intimidating. I like for people to also consider Julia the general purpose language that it is.

Also in the blog, e.g. “HackerRank CEO Vivek Ravisankar Says Julia’s Dynamism and Speed Makes Julia Better for High-Performance Machine Learning” and:

There’s also been a discussion about using Lua along with Julia.

News to me.


Not sure if it is explained but it is spelled out in the Introduction: Functional Mock-up Units (FMUs)

Regarding FMU: Perhaps an oversight, I would talk to the authors. A bit weird to have a discussion about here before that.

In general the community appreciates spelling out of ambiguous acronyms, see also our package naming guidelines, literally:


I know you didn’t write it but this is generally bad form. If you do define an acronym in the abstract also define it in the actual paper.

Actually it was defined in the Introduction but not in the Abstract. It should have been in both.


Hi all, I just came across this post and wanted to let you know that I am one of the authors of the linked paper.

You are right, the acronym should have been spelled out in the abstract as well, like it is in the introduction. The paper was targeted at the Modelica Conference 2021 and the acronym FMU (functional mock-up unit) is very well known there, because the Modelica Association is developer and maintainer of the FMI-standard.

A very short overview (and further literature) about what FMI and a FMU is, can be found in the paper section 1.2.

Thank you for your feedback. In future, feel free to also inform as directly, because arxiv supports versioning of papers, we can enhance the paper quality on basis of your feedback. We are always happy about any feedback we can get :slight_smile: