I’m proud to announce Kdomskia! :rocket: A new fr...

# feed

[Felipe Almeida dos Santos]

I’m proud to announce Kdomskia! 🚀

A new framework built on Compose Multiplatform that brings the power of HTML, CSS, and JS into Compose Web.
> NO Wasm or Canvas!!
Repo: https://github.com/kdomskia/kdomskia

Supported Platforms:
✅ Web (DOM) – with Compose HTML, Kobweb, and Beercss
✅ Android – with Compose Multiplatform
✅ iOS – with Compose Multiplatform
✅ Desktop (Windows, macOS, Linux via JVM) – with Compose Multiplatform

Check out a fullstack production sample built with Kdomskia:
Web App: https://pesto.kdomskia.io
Repo: http://github.com/kdomskia/sample-pesto

Feedback and contributions are welcome, I hope the community can help grow this together.

[Oliver.O]

Absolutely interesting, I need to look into that! Your next roadmap item is testing? If you want a fresh approach on that, you can check out testballoon TestBalloon.

[אליהו הדס]

Awesome 🤩

[Felipe Almeida dos Santos]

Hi @Oliver.O
Sure, the next thing is to add testing
TestBalloon looks promising, I'll check it out

[Oliver.O]

There is so much propulsion in the Kotlin ecosphere right now. Just learned more about Kobweb at KotlinConf 25, now seeing Kdomskia integrating everything and rising things to the next flight level. So cool! 🤩 The Cross-Platform Layout Guide looks particularly valuable, curious how that will work out when I get a chance (time) to try it.

Wouldn't adding a Wasm/JS target (still without Skia+Canvas) actually be a good thing, speeding things up for some use cases?

[Felipe]

Hi @Oliver.O
Sure, Kotlin is evolving fast and Kdomskia would not be possible without Kobweb
Regarding Wasm, before moving forward with this, I think we should first make a POC to check if there's a significant performance gain

[Oliver.O]

Evaluation of effort and effects makes sense. Adding the target may actually not be a big deal, according to my experiences while tracking Compose for Web on Wasm and JS basically since its inception (see my compose-counting repo).

My expectation would be that as long as there are few (HTML+CSS) elements Compose has to drive, you'll probably not see a big difference in rendering performance, and the browser's layout/rendering engine will absorb most of the load. I'm not sure where we'll see something like high-frequency animation-like updates driven by Compose. Maybe when moving lazy list objects around? On the other hand, Wasm code can be compact and may be interesting for load size optimizations.

Mileage will vary and checking proper use cases will help.

[Robert Jaros]

@Felipe I wouldn't want to diminish Kobweb's value (it's a great project) and I don't know how much Kobweb specific API is used, but for wasm target you could consider #C06UAH52PA7 as an alternative "engine" 🙂