Rust + Webassembly got me hype for a new slim-OS future. OS design

This is why we cant have nice things.

I have a feel that you're memeing. wasm is in very early stages so there's nothing to get hyped about yet

yes anon, i'm getting good feels about this one too as i graduate next year and this is exactly what i want to get into. Progressive Web Apps is a step in the right direction, wasm pretty much just seals the deal.

>>62446922
>OS design could become just making a nice browser
What a shit future we are coming in.

>>62446958
>>62447034
is it somehow a fundementally bad idea to have applications just work at almost native speed, regardless of what OS you're running? or are you just doing the classic /g/ neckbeard thing whenever something new comes up on the horizon?

>>62446922
>>62447103
>>62447019
>is it somehow a fundementally bad idea to have applications just work at almost native speed, regardless of what OS you're running?

With messages like that it might be easy to forget what shitloads of APIs a browser must implement, load all at once and then provide in a crippled, limited way to the actual code.
There is nothing slim, "just making a nice browser", "almost native speed" and "JVM but better" about a browser in many ways, and it always will be, even if they enable multithreading and direct DOM access in WASM.

>>62446983
no memes man, this is very exciting.

>>62447103
this person gets it. platform-agnostic near-native performance. how is that not exciting for everyone? not only rust, but it means that system-level languages will start gaining traction again for new problem sets.

>>62447207
you're conflating the idea of a browser with what browsers currently are.

with wasm there can be browsers that are written that can't even display basic webpages, but instead act only as a graphics context for a wasm file.

>>62447239
>platform-agnostic near-native performance. how is that not exciting for everyone?
see >>62447207


>but it means that system-level languages will start gaining traction again for new problem sets.
They don't.
It will be exactly the old problems, just shoved into other runtime environments.

>>62447275
>with wasm there can be browsers that are written that can't even display basic webpages, but instead act only as a graphics context for a wasm file.
It's called interpreter, anon.

File: comfy inside.jpg (63KB, 1280x720px) Image search: [Google]

63KB, 1280x720px

>>62446922
wake me up when a browser can render apps without needing 200mb+ per page.

>>62447350
make that 400, the manufacturers are pushing

>>62447103
>have applications just work at almost native speed, regardless of what OS you're running
isn't that exactly what java promised in the 2000s ?
how did it turn out ?

File: screenshot-20170723180632.png (89KB, 1295x733px) Image search: [Google]

89KB, 1295x733px

>>62447239
>platform-agnostic near-native performance
https://github.com/graalvm/sulong

>>62447452
pretty great desu

Brendan Eich is a complete nutcase

>>62447335

I think you're missing the point senpai.

>>62447483
>https://github.com/graalvm/sulong
>JRuby
>best specialized competition

>>62447483
>uses an oracle source to back java performance claims

>>62447502
I'm not, stop being fucking vague.
If you want a widely-usable standard that lets you do proper graphical programming, you surely won't get it from W3C and apologists, because they failed at so many APIs, just don't expect too much from them.

I agree it's high time people gather and create a good, FOSS, multiplatform GUI library that stands between Tk and Qt that works natively in [my favorite language] and not in [your favorite language], but considering costs and market I don't see that happening any time soon.

>>62446922
Rip GPL. Hello copyleft.

File: 14888055594900.jpg (692KB, 2560x1600px) Image search: [Google]

692KB, 2560x1600px

>>62447103
>at almost native speed
Thank god we have all dem megahertz and megabytes to handle "almost native speed" so end-user wouldn't notice such nonsence.

>>62447515
is there a faster a faster ruby implementation (except truffleruby) than jruby? I'm not really asking because the answer is no

>>62447568
it's oracle labs, not oracle. they're researchers and they don't sell any products. they have literally no reason to lie

btw you also get full language interoperability with no overhead (the language has to run on graalvm though). which means you can literally call R or Ruby functions from C code

>>62447603
>proper graphics and GUI programming
fixed

>>62447630
>is there a faster a faster ruby implementation (except truffleruby) than jruby?
Pretty sure there is still Rubinius.

>>62447630
>they have literally no reason to lie
except for the part that they're oracle employees ?
they don't even have to lie, they just have to produce shitty "native" code

I remember when java was supposedly as fast as c/c++ code... with no optimisations enabled

>>62447658
Nvm, they seem to have broken it.
Still, JRuby isn't that much of a win, is it?

>>62447627
wasm is barely even out and yes, even now, in most cases the difference in performance is barely noticeable at all. if you are really sperging out over fractions of seconds, then don't worry, the standard is still new and there's plenty of room for optimisation.

File: rubykon_ipm_new1.png (18KB, 797x530px) Image search: [Google]

18KB, 797x530px

>>62447568
>>62447630
also their "java performance claim" is that it runs 1.02x faster on graalvm than on hotspot which is not really that hard to believe

>>62447658
>faster

>>62447706
see >>62447696
However, that apparently is just temporary, since they disabled the JIT compiler to switch it.

File: WebAssembly.png (1MB, 2087x1329px) Image search: [Google]

1MB, 2087x1329px

>>62446922
webassembly is slower than java
http://www.stefankrause.net/wp/?p=405

>>62446922
I couldn't care less as long as noscript and umatrix continue keeping that trash off my computer.

>>62447843
we couldn't care less about opinions of autists on this subject

>>62447917
Then why are you shilling it on /g/? Go shill on Facebook or LinkedIn then

>>62446922
>OS design could become just making a nice browser
>implying that's not the case already

smart phones do this already, outside /g/ no one cares if you use a macbook or a chromebook as your facebook machine
>what's with all the *books?

>>62448032
>anonymous is one person
welcome newfriend, please read our new user guide :)

saying "hey, lets have a scripting interpreter so we can have active content on web pages!" is the original sin of the web. Its something that we should never have done, and that we should have known we should never do. But we did it, and now we are cursed to suffer at the hands of this utterly terrible idea for all time.

>>62448108
>The pronoun you is the second-person personal pronoun, both singular and plural, and both nominative and oblique case in Modern English.

>>62448460
this

>>62447816
What kind of shit nbody benchmark completes in only a few seconds? You don't even start running into the memory issues

>>62448654
https://benchmarksgame.alioth.debian.org/u64q/performance.php?test=nbody
this one

>>62448652
>The pronoun you is the second-person personal pronoun, both singular and plural, and both nominative and oblique case in Modern English.
what does the autist mean by this? how am I shilling for webassembly exactly?

>>62448460
given the opportunity to redesign the web from scratch, what would be the logical way to go about it? i've read about this years ago but i forget what kind of an implementation it was referring to.

>>62448460
>saying "hey, lets have a scripting interpreter so we can have active content on web pages!" is the original sin of the web.
it's what allowed the web to become so powerful in the last decade

>>62449593
Yes, the web becoming so powerful is one of the many awful consequences.

>>62449623
yeah, it's so fucking awful to be able to load new posts on 4chan without reloading the whole fucking page or to follow quotelinks without jumping up and down the thread like a retard