Wasm will fix the web.

>>58617418
No, it won't.
Web applications will remain bloated because incompetent developers can write inefficient code in any language.
It will just become more proprietary.

>>58617418
so web pages will become binary blobs now?

Not really.
You still need JS to interact with the page in any meaningful way.

web assembly can only export functions that can do calculations more efficiently than JS.
It can't react to events or trigger any sort of reaction.

Does anybody else have the feeling that wasm is intended to replace the JVM?

I mean the patents on Java have got to be ready to expire. It is a stack based machine (like the JVM). Oracle has gotten pissy with people about Java in recent memory.

The effect of web assembly is just what java applets were supposed to be: platform independent code running in the browser, only web assembly is going to be able to tie into the browser much more deeply (as soon as they figure that part out)

All the major consumer software vendors are slowly giving up on selling software anymore. OSS has eaten their lunch. It is software as a service / subscription from now on. With Oracle waffling on Java a new platform to build on is needed.

Given the platform can run at native speed, access the gpu (webgl), etc. doesn't this become the perfect platform?

>>58617476
>It will just become more proprietary.
Proprietary to whom?

>>58617691
wasm is in the client though. JVM is mainly used in the server.

is wasm like llvm where languages are built on top of it?

>>58617418
>putting webdevs anywhere near low level code
I wonder how this will play out

>>58617691
modern browsers are definitely shaping up into full-blown JVM/CLR tier runtimes. with W3C standards you don't have to worry about your environment and/or libraries being multiplatform either, every single OS worth using has a modern browser available for it.

hell, at work we're already planning on moving all our ERP software to web applications.

>>58618217
>hell, at work we're already planning on moving all our ERP software to web applications.
Is there any reason for doing that other than because it's cool and hip now to have web apps over the real thing?

>>58618252
>Access the same service on different devices
Gee, I wonder if there is a reason.

>>58618252

Deploy over the web means everything is up to date, you get runtime statistics from users (no choice about it), and if you keep functionality on the server, piracy is impossible.

I assume it will suck for the users.

>>58618252
Like you do realize why it's "the cool and hip" thing now? Or are you seriously this blind to accessibility and believe every single change that occurs must be some bandwagon bullshit?

>>58617418
Web assembly is the death of the open web.

Say goodbye to javascript and html. Pages will ship web assembly instead. The ad industry will use it to hid ads. And since most of them won't provide source code, you will just be running closed source apps.

It's the dream of Java. Make every web a java applet. It fucking sucks.

I thought it was a /g/ meme. Holy shit.

>>58617418
No, JS is just one small part of the problem that is the browser.

>>58618725
Sadly it's not nearly as interesting as the name suggests. WebAssembly is basically just another javascript with arguably better syntax. It's actually a kind of bytecode similar to the code that runs in JVM

File: urielwasright.png (26KB, 820x1172px) Image search: [Google]

26KB, 820x1172px

this is literally the most cancerous piece of technology I've ever seen.

at this rate, 9front will be the last man standing in the OS wars

>>58618904
The only thing he was right about was the decision to kill himself.

>>58618147
>wasm is in the client though. JVM is mainly used in the server.
Basically browsers have already superseded the JVM, except they use javascript source instead of bytecode, so they're usually not as efficient.

So this is just a new runtime that will happen to run in browsers, but could just as easily be packaged on its own. And ideally it will end up supporting a bunch of languages. So devs can make cross platform programs without the need for Java/JVM-based languages.

>>58618279

Stallman must be shitting bricks at this trend

>>58617418

Can anybody give an example how this stuff looks like?

>>58619090
Like assembly code. Just that it runs in the browser.

The idea is you'd compile a higher level language into it. I believe there's already a compiler for C++ and Rust.

>>58619116
neither C++ or Rust are anywhere close to usable. It's in the "tech demo" phase.

>>58619116

I know what ASM is and I know what JavaScript is. I just don't get how you can from those technologies into one?

Do you have special flags for "window.onload" or what?

>>58619147

>C++
>not usable

Hahaha, what?

>>58619190
The C++ compiler for wasm you context-free retard. It's missing a lot of important things that cause a lot of existing applications to NOT compile.

>>58619215

OK.

So basically it's about an elaborated compiler and not about writing a website in ASM?

>>58619231
It's about creating a standard bytecode that all browsers can read, and compilers for many different languages so that they can all be compiled for the browser ie all used for frontend web dev.

>>58619231
> an elaborated compiler

No, it's an LLVM backend that translate LLVM-IR to web assembly. It's no different than any other assembly platform. The problem is not all features are finished yet, so theres no way to represent more complicated parts of C++'s multi-threaded memory model in the browser without support in the browser to run it.

>>58619215
>you context-free retard
10/10

One way to look at it is that it's basically ahead-of-time compilation for web software. Previously software either had to be compiled at run time (just-in-time) or it could be "compiled" kinda sorta ahead-of-time with asm.js or by prediction in the browser to figure out what scripts are the most costly performance-wise.

Webasm allows web devs to "compile" their software from one high level programming language to a standard IR that will be executed more efficiently than it would be if it were converted to javascript or even asm.js

File: code1.png (154KB, 834x415px) Image search: [Google]

154KB, 834x415px

>>58617505
It's not like current proprietary JS isn't some form of blob. Pic related is what most proprietary JS code looks like already. It's still source code, but no one will ever be able to read this

I don't see the problem, if you don't like Javascript there are a shittone of languages that compile to it:

Scala
C#
Python
Java
C
Ruby
Erlang
Elixir
Scheme
Clojure
D
F#
OCaml
Haskell
Go

Also you have supersets of JavaScript like Typescript and whatever coffescript is.

>>58619393
But it's generally easier to de-minify javascript than it is to decompile assembly.

>>58619459
If WebAssembly is similar to the CLR (or JVM), decompiling it should give reasonable results. You can decompile most compiled CLR libraries back to somewhat readable C# code for example.

>>58619336
IMO it may also allow to do some premature optimization like LTO, dead code removal or stop that stupid implicit type conversion (that causes more problems than benefits). I'd like to see some language that has JS/Java-like simple syntax for newbies, compiles to wasm and performs a lot of optimizations like C/llvm does.

File: Chomsky_2004_grin.jpg (1MB, 1712x2288px) Image search: [Google]

1MB, 1712x2288px

>>58619215
>context free retard

>>58619090
>example how this stuff looks like?

Tt's a compilation target for a new, common virtual machine that has hooks into a pages DOM and whatever else.
i.e., you write your code in one of any number of higher level languages like C++11, C#, etc., and compile it into WA instead of x86_64, ARM, or JVM.

The goal is to enable a common platform without any of JS's retardation or having anything to do with Oracle/Java.

>>58619449
because it'll make everyone's browsers and web services seem at least as good if not faster while costing companies less to host content.

Translating binary VM code will always be faster than lexing, parsing, and doing semantic analysis for human readable code whenever the same middle/back ends of a compiler are being used.

>>58619593
how long have you been waiting to use that reaction image?

>>58619449
i invite you to crosscompile C++, C, or hell even Python to JS and see the pitfalls of doing it for yourself. Not to count the shit ton of hacks you have to do be able to manipulate the DOM in the same way as you would with JS.

p.s. i hate javascript with a burning passion, so this isnt a js pajeet meme.

WASM is the future. if you wrestle with wasm adoption, prepare for bleak future career opportunities

>>58617418
So I'm a C++ dev and I'm at least curious about WebAssembly.

What sort of APIs and threading models will it provide applications to interact with?

>>58617418
No.

>>58623462

> t. JS "Dozens of frameworks" Fag