> Probably because many are purists. It is like how anything about improving Electron devolves into "you shouldn't use Electron."

The Electron debate isn't about details purism, the Electron debate is about the foundation being a pile of steaming dung.

Electron is fine for prototyping, don't get me wrong. It's an easy and fast way to ship an application, cross-platform, with minimal effort and use (almost) all features a native app can, without things like CORS, permission popups, browser extensions or god knows what else getting in your way.

But it should always be a prototype and eventually be shifted to native applications because in the end, unlike Internet Explorer in its heyday which you could trivially embed as ActiveX and it wouldn't lead to resource gobbling, if you now have ten apps consuming 1GB RAM each just for the Electron base to run, now the user runs out of memory because it's like PHP - nothing is shared.

Or these devs & users can migrate to a PWA. Which will have vastly less overhead. Because it is shared, because each of those 10 apps you mention would be (or could be, if they have ok data architecture) tiny.

> Or these devs & users can migrate to a PWA

PWAs have the problem that for every interaction with the "real world" they need browser approval. While that is for a good reason, it also messes with the expectations of the user, and some stuff such as unrestricted access to the file system isn't available to web apps at all.

Removing layers is hard though, better to have electron host a WASM application which will become a new "native" that gets argued semantically.

Each person seems to have their own bugbear about Electron but I really doubt improving Electron to have shared instances a la WebView2 would make the much of a dent in the hate for it here.