Dukat is interesting. Right now, typescript is a popular replacement for javascript. However, I've always regarded it as a bit of a gateway drug for people to find their way to a bit more higher level languages. Like Kotlin.
What Dukat does is leverage typescript type definitions for javascript libraries to automatically generate Kotlin type wrappers. This basically means kotlin is inching closer towards the goal of being a replacement for typescript/javascript. Right now it's not for the faint hearted as the tooling is still a bit lacking. However, the ergonomics of using a proper IDE to do web development is very nice and Kotlin comes with a lot of language features that are very nice to have when doing complex asynchronous processing in modern UIs. E.g. co-routines vs. async-await in typescript is just a lot less clumsy/painful/error prone. Dukat is of course experimental.
I'm not actually sure that playing nice with the javascript world is relevant long time because of other things happening in Kotlin that will likely make that less important. I expect a lot of things will start changing during the 1.4 and 1.5 cycle in the Kotling world:
1. kotlin-js tooling will stabilize and become more robust (much needed). Right now if you really want to, you can do some nice frontend development with just kotlin. However, my experience is that you end up fighting a lot with tools and frameworks and the general lack of documentation for all of this. I can't recommend it for that reason. However, if you can deal with that, there are some clear advantages like having a proper IDE with advanced refactoring, autocomplete, and other modern features. Additionally the Kotlin language has a lot of nice features relative to typescript.
2. kotlin native will come out of beta; probably with 1.4. It's very clear that Jetbrains is trying to establish Kotlin as a solid cross platform language for mobile development. Apple is pursuing a similar strategy with Swift.
3. Multi platform kotlin libraries are about to become a thing. These are libraries that look and feel the same whether you use the jvm, js, or native compiler and wrap/proxy platform specific functionality where appropriate. Having multi platform libraries for things like dealing with IO, Json serialization, HTTP, etc. means you can have cross platform kotlin code bases that compile for multiple platforms. Currently this stuff is a bit in flux but when this stabilizes, I expect a lot more libraries to start popping up on Github.
4. Kotlin native also means WASM support is going to be there as well (basically Kotlin native uses llvm and can target anything llvm can target). With WASI and other interesting things happening in that space and e.g. things like garbage collection inching closer to being standardized in browsers, Kotlin becomes a solid language for targeting this as well and maybe a nice alternative for people for whom rust is a bit too alien/hard right now.
5) All of the above means that Kotlin is going to start becoming a more viable choice in a lot of places where it is currently not an obvious choice: server-less computing, browsers, desktop software, command line tooling, etc. Also it means attention will shift to doing things the Kotlin way.
So, while Dukat is nice for integrating legacy typescript libraries, it's going to be more interesting to utilize multi platform kotlin libraries instead. Also transpiling to js becomes kind of unnecessary if you don't need javascript libraries and can use WASM instead via kotlin-native.
What Dukat does is leverage typescript type definitions for javascript libraries to automatically generate Kotlin type wrappers. This basically means kotlin is inching closer towards the goal of being a replacement for typescript/javascript. Right now it's not for the faint hearted as the tooling is still a bit lacking. However, the ergonomics of using a proper IDE to do web development is very nice and Kotlin comes with a lot of language features that are very nice to have when doing complex asynchronous processing in modern UIs. E.g. co-routines vs. async-await in typescript is just a lot less clumsy/painful/error prone. Dukat is of course experimental.
I'm not actually sure that playing nice with the javascript world is relevant long time because of other things happening in Kotlin that will likely make that less important. I expect a lot of things will start changing during the 1.4 and 1.5 cycle in the Kotling world:
1. kotlin-js tooling will stabilize and become more robust (much needed). Right now if you really want to, you can do some nice frontend development with just kotlin. However, my experience is that you end up fighting a lot with tools and frameworks and the general lack of documentation for all of this. I can't recommend it for that reason. However, if you can deal with that, there are some clear advantages like having a proper IDE with advanced refactoring, autocomplete, and other modern features. Additionally the Kotlin language has a lot of nice features relative to typescript.
2. kotlin native will come out of beta; probably with 1.4. It's very clear that Jetbrains is trying to establish Kotlin as a solid cross platform language for mobile development. Apple is pursuing a similar strategy with Swift.
3. Multi platform kotlin libraries are about to become a thing. These are libraries that look and feel the same whether you use the jvm, js, or native compiler and wrap/proxy platform specific functionality where appropriate. Having multi platform libraries for things like dealing with IO, Json serialization, HTTP, etc. means you can have cross platform kotlin code bases that compile for multiple platforms. Currently this stuff is a bit in flux but when this stabilizes, I expect a lot more libraries to start popping up on Github.
4. Kotlin native also means WASM support is going to be there as well (basically Kotlin native uses llvm and can target anything llvm can target). With WASI and other interesting things happening in that space and e.g. things like garbage collection inching closer to being standardized in browsers, Kotlin becomes a solid language for targeting this as well and maybe a nice alternative for people for whom rust is a bit too alien/hard right now.
5) All of the above means that Kotlin is going to start becoming a more viable choice in a lot of places where it is currently not an obvious choice: server-less computing, browsers, desktop software, command line tooling, etc. Also it means attention will shift to doing things the Kotlin way.
So, while Dukat is nice for integrating legacy typescript libraries, it's going to be more interesting to utilize multi platform kotlin libraries instead. Also transpiling to js becomes kind of unnecessary if you don't need javascript libraries and can use WASM instead via kotlin-native.