/tech/ - Technology★

No it's literally javascript. Web Assembly is where you take other languages and transpile them into javascript.

>>854178

WebAssembly is a binary format, executed within a sandbox.

>>854182

The way i understand it a fundamental difference to JS is the ability to compile line by line, to make use of streaming compilation like in the linked article.

Wouldn't the same thing be able with an interpreted language like Python?

>>854182

No you're thinking of some vaproware that will never make it into mainstream use. The WebAssembly you're going to fucking GET is asm.js.

>>854184

asm.js is not WebAssembly, they are two different things.

>>854183

Python is interpreted, not compiled. This "streaming compilation" isn't like interpreting, it just means compilation happens without requiring the entire source file at once.

>>854185

Yes, one is vaporware that will never see mainstream use and one is the one you're going to fucking get.

>>854190

Why would you start talking about asm.js in a WebAssembly thread?

In spite of its name, looks like yet another "bytecode" platform like .NET, JVM, or P-System. From what I can tell, the main advantage over asm.js is just that it's built into the browser as a static library, so the bytecode VM boilerplate doesn't have to be downloaded with every single thing that uses it.

I've never understood the attraction in this trash. Why not just distribute compiled binaries for an actual ISA like x86 and ARM? On the actual hardware (which it would be 99.999% of the time) it would be a zillion times faster and lighter, it would be just as secure run inside a VM, you could still do dynarec just like LLVM does if the binary was really old, and on the 0.001% of edge cases where you were using a different ISA like MIPS/PPC/etc it would emulate/recompile just as fast as a fake "bytecode" ISA.

>>854199

How about we replace javascript with python or something else less shit, and then stop trying to make serious programs in the fucking web browser.

>>854199

>it would be just as secure run inside a VM

Running random assembly from the web would not at all be secure.

There was never a worse time to roll out native code in a browser than right now. Even lowering the precision of performance.now() won't mitigate the kind of attacks that are feasible through spectre and meltdown.

>>854182

This is not going to end well.

>>854201

>stop trying to make serious programs in the fucking web browser.

Look dude, the only way this is gonna happen is if there is NO programming/scripting language AT ALL.

If you replace javascript with python, people will just figure out how to do the same shit with python.

We have to just pull the plug on the whole thing if we want change.

Web Assembly is just going to enable more proprietary code to be developed.

I think it's a mistake. An HTML document viewer does not need to execute arbitrary code downloaded over the internet without even asking the user.

>>854212

I'd ban you just for your fucking post formatting. Kill yourself.

>>854213

>we have to just pull the plug

>we

What makes you think you have any power or say in the matter?

>>854215

Web browsers are no longer "HTML document viewers", buddy.

The problem with WebAssembly is that it may be the end of adblocking as we know it. Since it's a binary, adblockers won't be able to approach them as they approach JS.

>>854210

Everyone loved java applets right? :^)

As long as webassembly doesn't have access to the DOM, it should be as easy to block as flash, right?

>>854304

>2019

>new pages are full of nothing but javascript callbacks with webassembly "backend"

>>854296

I don't. What I'm referring to will never actually happen, but that's what would have to happen to get people to stop making programs in the browser. Make it so that they can't.

It's like

>>854215

and

>>854297

are talking about. You said earlier that people shout stop trying to make serious programs in the browser, but that ship has already sailed, sadly.

>>854303

holy shit I remember those!

>>854297

Yeah they is, except sometimes the document isn't static. But if they do want client-side scripting, fuck them. Homie don't play dat. Only Lynx and Links -g muthafuckas!

>>854325

Javascript is more ethical than black box server-side logic (barring honest use of the AGPL).

It's just huge unreadable packages of minified javascript (and now bytecode) that subvert that.

>>854326

Nah, I don't really care what they run on their server. That's their hardware. It only needs to be open source to run on my gear.

But anyway I don't want to scripts running in my browser anyway, because it's just one more attack vector.

>>854327

The source is right there retard, it is sent to you.

>>854326

>What does /tech/ think about Web Assembly?

It's reinforces DRMs and botnet.

>>854183

first need into strong typing

>>854300

couldn't you just choose not to execute any binaries, producing more or less the same inconveniences as with not executing JS?

>>854389

Perhaps the main reason JS isn't even more prevalent than it already is, are performance issues.

Take these away, and Internet 3.0 4.0? whatever marketing jews might call it might get rid of HTML entirely, with all but legacy sites simply serving one binary as a site.

>>854395

This is a wet dream for DRM. Software as a service perfected. You don't just pay for it, you don't just get a license to use it, you don't even get to have it on your hard drive. It's software STREAMING.

WebAssembly theoretically is a way to write the equivalent of JS functions but faster. It does this by letting the user write byte code that then gets compiled and executed in the same sandbox as JS code.

The biggest problem is that you have a good idea of what actual CPU instructions WebAssembly byte code will map to, which makes it easier to exploit CPU bugs. And like >>854208 said this is bad timing given the recent CPU bug disclosures.

>>854178

>>854184

>itt retards

webassembly exists today and works by default in all modern browsers.

it is a bytecode that gets directly jitted to the native CPU opcodes by the browser without needing to go through are the parsing and optimization steps that javascript does to get turned into opcodes.

>>854177 (OP)

I think its got a way to go before it becomes useful.

Threading and DOM/Web API access is essential, while its "being worked on," who knows if it will ever see the light of day

>>854187

There are several ways to compile Python.

As a matter of course you should be disabling this cancer.

https://github.com/stevespringett/disable-webassembly

>>854177 (OP)

It's a transparent power grab, the idea here is to eliminate ad blockers and increase tracking. I feel they will slide this in by requiring it for streaming video, integrated with DRM, think Netflix and Amazon.

For your safety of course.

>>854793

While it will certainly make adblocking more difficult, it won't eliminate it. The adblocking community has shown to be very driven to their cause, even going as far as to create anti-anti-adblocking userscripts.

wasm bytecode can still be intercepted, dissasembled, and modified. Its hardly any different than the obfuscated javascript used today to serve ads and block adblockers. If that really is the subversive goal of wasm, it will be a failure.

>>854829

Why go through all that trouble though? At some point you have to just give them the finger, instead of keep piling on tons of hacks to play their stupid game. No fucking social media site or video streaming or gaming shit is worth that kind of headaches.

>>854177 (OP)

Checked em.

>coding is hard

>making coding for niggers

Terry was right

What's not to like? It allows me to avoid JS and write programs in a sane language like C++ and then having it run on all platforms without cross-compiling anything which is a huge win.

My only problem is that performance isn't exactly close to native despite what they advertise and it's a bit rough around the edges as of now.

>>854998

And a real l33t h4xx0r like yourself uses?

it wasn't designed in a weekend so it's instantaneously better