Replying to Avatar jack (n/acc)

That's probably the toughest part

The advantage that I see of this method though, is that the project in question is already an abstraction over Swift UI, gtk3, and whatever Windows has going on so you could make an abstraction over egui using the same technique and then load it over the wire with nostr events

and then after you do that, you only have to ship one version of your product. the client library just adapts that abstraction.

lots of ideas here, many ways to explore it.

curious to see what you come up with though
Avatar
jb55 1y ago

i think I would want a very small core for the ui, like something at the level of webgpu command buffers. Then of course you could provide higher api levels so you don’t have to target the gpu directly, but that would always be available if you wanted to build a game or egui ui or something

Reply to this note

Please Login to reply.

Discussion

Avatar
jack (n/acc) 1y ago

I feel like in a lot of cases though that could lead to websites shipping their own UI frameworks (egui, imgui, nuklear), which is not ideal.

I think the advantage here is that it picks the default browser native framework automatically, and then you could probably give it some kind of GPU access to.

Avatar
jb55 1y ago

why is it not ideal? there could be a default framework but you don't have to use it.