<https://9to5google.com/2018/12/11/aosp-fuchsia-sd...

# random

[itnoles]

https://9to5google.com/2018/12/11/aosp-fuchsia-sdk-device/ Oh Boy

[sksk]

ping @gildor you see, i was right, i said that since last year, i even suggested to work on fuchsia target for KN 😛, i'm pretty sure Google want to get away with their decade old code base full of Java code and start fresh with Dart/Swift, they even considered Rust, but then the guy responsible of that left google, maybe they'll focus on Swift/Dart for iOS compatibility

[sksk]

kotlin should do something there imo

[gildor]

i’m pretty sure Google want to get away with their decade old code base full of Java code and start fresh with Dart/Swift

Ohh, common, really? How could you did such conclusion from this article?
Swift, really? Even this is so funny that not even sure how to comment

but then the guy responsible of that left google

Exactly! That what I’m saying! There is no Google, there are some particular teams, particular PMs and even developers which can quit or just fail. And this happening all the time, and very few products, even from goodle are successful)
Your message reminds me 2 episodes:
1. Beta of Google Wave - “Ohh! Google going to replace emails and messaging with a new revolutionary product! Give me invite”. Closed in a few months, teams continue work on another departments
2. Commit to Swift to support Android NDK - “Ohh! Google going to replace Java with Swift”. 2 years after - Nothing. Swift is too heavy weight runtime, not backward compatible, has no tooling support etc and has very limited usage outside of iOS

if seriously “suggested to work on fuchsia target for KN”. Why? Why someone should spend time and money for development for non released project and completely unknown goals, SDKs, supported platform (at least without moneys from OS manufacturer or some real internal information, not from commits that do nothing)
There is a bunch of things to do, for example support platforms with 2 more billions of users than Fuchsia and with thousands of device models more, or improve K/N multithreading, or improve K/N tetris demo %)
Literally any other thing would be better to do than that.

Is you so sure that would be successful, maybe you should start own work on it and invest time? If you so sure that it’s gold mine, next big OS and so on.
Also, if Fuchsia supports llvm and posix what exactly you expect to support?

Again, maybe you right, but I would like to see your article with some tech details, rather than this click bate article with no information, full of speculation and wrong statements

[sksk]

you sounds really on the defensive about it, dunno why ?

if seriously “suggested to work on fuchsia target for KN”. Why? Why someone should spend time and money for development for non released project and completely unknown goals, SDKs, supported platform (at least without moneys from OS manufacturer or some real internal information, not from commits that do nothing)

Should learn about R&D and why anticipating is more important than just reacting about new stuff, that's how you get a lead, and that how you prove other people (future consumers/users) that you care about what you do, and the environment everyone's in

Jetbrain is a business, there is lot of buzz about fuschia, completely ignoring it is plain stupid, Apple and Swift got huge press coverage when a PR was sent to support that language

Same for webassembly, huge buzz about it and yet Jetbrains doesn't try jump into it more publicly

[sksk]

Mozilla got it right with Rust, that's how you hype developers, and that's how companies end up using that product, because these developers recommended it

[gildor]

Should learn about R&D

R&D for another non-finished R&D? Meta-RnD

Should learn about R&D and why anticipating is more important than just reacting about new stuff

This is not about reacting on new stuff in general.
This is a very particular case.
Some team in Google work on the completely new OS. This OS is not released not even in the status of technology preview or pre-alpha, just nothing yet, we don’t know about particular plans.

that you care about what you do, and the environment everyone’s in

What exactly do you mean? Are you talking about care for people who care about Fuschia? Isn’t this super limited auditory?
You also, for example, can do good things for much more developers: improve tooling, improve K/N for existing platforms or just do compilation faster

there is lot of buzz about fuschia

Exactly, there is a lot of buzz, it’s good for tech websites because they need something to write about each day, so they write about each meaningless commit to Fushcia.
Of course, it’s an interesting topic, some new OS from Google, of course by itself it’s interesting.
But you say not just keep eye on it, you talking about a thing that replaces years and years of development, one of the most popular platforms in history (JVM) and OS with billions of people (Android).
And I don’t think that this is somehow close to reality, looks that you just under buzz effect rather than have real arguments.
What it will be? Who knows. Maybe Google wants to escape from Linux and have own platform. Maybe they want to refresh Android, but Android is a huge platform and OS is not the only thing.

completely ignoring it is plain stupid

What “ignoring” means? Do not read news about it? Ignore release notes when they appear? I don’t think that anyone in our area like that, including me. I keep eyes on it. And for now we have a few things to discuss, but nothing to do (like write code, release support). Just wait and see what is going.

[gildor]

Apple and Swift got huge press coverage when a PR was sent to support that language

Exactly, this is just press (or more precisely just some tech news websites). And what? It’s still nothing as before. Swift compiles using LLVM, and K/N does the same, it’s not a big thing, just compilation for one more OS (but in this case not release with 0 users and very few devs). But the compilation of language for particular OS is not enough, you need also platform libs, bindings for API, everything what is required for real native development and what is not exist for Fuschia, or at least just some preview prototype.
Also, just a side note, I don’t think that Swift will be bundled to Fuschia as some core technology, just because it’s not stable enough, you don’t want to rewrite OS and all programs just because Apple broke backward compatibility, yeah I know, they promise to stop do this, but I would be really careful about it

Mozilla got it right with Rust

Do you really compare Rust in this case? Rust is a project started by one of Mozilla developer so in this case Mozilla and Rust case much closer to JB and Kotlin, than to [ANY_COMPANY_EXCEPT_GOOGLE] and Fuschia
Also, when Fuschia will be on the same level of maturity, I will be fine with it. Rust is a project with 12 years history and almost 4 years after 1.0

and that’s how companies end up using that product because these developers recommended it

And this is exactly what happened with Kotlin, so for me, it’s an argument for existing Kotlin strategy, rather than about “let’s do hype driven development for our language”

[gildor]

You all the time talk about “support language in OS”, but it’s not what modern OS should do. OS is a platform that should provide good HW abstraction layer and APIs which can be used to run your own platform or programming language.
You showing this article and said that Kotlin team should do something, but what if Kotlin should do nothing and Fuschia is just compatible with POSIX or they just natively support Dex (I remember also some commits about this), in this case, you should do nothing and just wai

Same for webassembly,

But it’s not the same.
K/N already have experimental support compilation to wasm using LLVM, is not what you expect?
Big difference comparing with Fuschia, is that wasm already available on the most popular application in the world: browsers
But even in this case, it’s not so simple, many pieces are not available. Some time ago I had a conversation with people from WebAssembly Working Group and they looking into future, not to current very basic implementation and always say “Do not say WebAssembly implementation, this is just an MVP”, not real platform yet under heavy development. To be really useful we waiting for GC, better JS interop, threads, tooling etc.
And even without a huge effort from JB we already can try it with K/N. But isn’t it obvious how different situations between technology that already available almost on every device and OS that has 0 consumer devices and has no release status