Great talk, @smith.adriane ! I've been building some personal desktop applications using membrane and I am feeling cautiously optimistic that the Functional UI Model will "win" in the end, in the sense that it will be a powerful and usable option for developers who seek it out, which many will. You had a comment at the end of the talk about people scratching their own itch to swap out various pieces of membrane, which could eventually leading to membrane proper eventually disappearing. I think that's a humble insight on your part, but it also made me feel hopeful about the approach.

Really enjoyed your London Clojurians talk @smith.adriane Really quite excellent.

me as well!

after thinking about it for a day I immediately thought of a project I recently started. A website that has a responsive hexagonal grid as its frontpage navigation. Like this (work in progress):

The hexes are fixed size. Centered with some flexible margin at the sides, grow infinitely downwards.

In imperative JS (I also used the bounding box API i think...) this is just a few lines. handling media query events (resize) and doing some simple modulo math etc. In pure CSS this is idiotic. AKA not declarative at all, messy etc. Looking at the code you wouldn't recognize in the slightest what it expresses...

Neither solution is really declarative, both are hyper specific. Needless to say there's no clear connection to what the designer wants to express with how I expressed it, so there's no way they could touch the code and play with it.

But on a conceptual level, the constraints are very clear, the relationships are pure. But with the tools I was using there's no way to express the essence of it.