MWS: Could we design MWS so it eventually becomes a local-first app?

[MagicJF]

I had an intuition about what I would like the final MWS product to be like but couldn't articulate it properly until I saw this Martin Kleppmann 's talk.

In case these principles are in the scope of MWS, it would be beneficial to have them present on the background when designing the sync and auth modules so in the future there will be no friction to evolve into a local-first MWS.

Regarding sync, I found this sync server's design talks very inspiring, probably we can adopt some ideas.

Linear's CTO talk 2024

Linear's CTO talk 2020

[Schmavery]

I'd also love this -- I've been using TiddlyPWA on my own wiki to allow me to use TW as a progressive web app, with local storage which is synced with the server when connected to the internet. That said, I've been really interested in the MWS development, especially around multi-user stuff, and it would feel a lot simpler to have the tiddler files viewable directly on disk somewhere (rather than being in the TiddlyPWA SQLite db) oops just realized that MWS also uses a sqlite db, nbd.

[Arlen22]

As far as whether it can be made "incredible-journey-proof" (love that name), I'm not really sure. Obviously that's the entire premise of TiddlyWiki as a whole, but MWS specifically is more about giving you the tools to host your own data. It's targeted primarily at use-cases where the server owner also owns the data, although we are certainly building it with the "public hosting" scenario in mind.

I think once we get some of the core features hammered out we'll be able to figure out how much we can do with local-first sync and stuff like that.