Obsidian is using electron, so the source is already somewhat available anyway. I understand them not making it open source, and risking someone forking it and harming their business. But considering the situation, I would think making it at least source available on a popular forge, where people can make issues and open merge-requests, might be a beneficial thing.

There are a bunch of small problems people encounter here and there, which usually will never be solved by the company. Giving the community a route to improve their tool, would be good.

Does anyone know if it's possible to have a core which is unsandboxed, but load plugins which are sandboxed? This seems like a great solution if so.

This is one of the main use cases for Webassembly outside of the browser.

I think we will soon see the ability to write plugins that can even run server-side of SaaS solutions.

The PKM I've been using lately, SiYuan, does exactly that, and I think their business model isn't bad: the client is fully FOSS, there are some client-side paid features with a one-time subscription (WebDAV/S3 sync "bring your own server") and some server-side paid features with a more expensive recurring subscription (cloud space provided by them).

I don't particularly like client-side paid features, but:

- The client is fully FOSS, you can just patch the license check out. In fact, there are some forks on GitHub that do just that and provide binaries, and the authors don't seem to care, they even acknowledged them on Twitter (https://x.com/b3logos/status/1928366043094724937).

- There are plugins to sync without a paid plan

This works out quite well for them: if you choose a fork or a sync plugin, you don't get the same support that paying users do, so many users still end up buying a license. But you don't need to, which makes the whole thing not user-hostile.

I have bought a one-time license myself, and I'm very happy that I'm supporting the development of a FOSS project.