/tech/ - Technology★

Fuck Javascript niggers. All hail Rob Pike.

>>926595 (OP)

Is that what Markiplier would look like if he consumed more soy?

>>926595 (OP)

>just import files from the internet

This is such a bad idea. Every time you start your program, you would have to wait for 50 MiB of dependencies to download. What if someone made an iseven dependencies, it get included in a big website. IsEven is then updated to allow for the creator to have remote execution and you get pwned. Additionally dependencies over http could be mitm for achieving rce.

>>926595 (OP)

>Typescript is the default language

>not Python 3

fuck this gay earth

>Node 2

Node.js is up to 10.3.0, he's living in the past

>>926642

He said the idea is to cache the files, just like a web browser would.

This is similar to how Go works, where you can import a git repository on the fly.

>>926659

Because it would mean actually learning a new language, meaning you can't take the same guy for frontend and backend, meaning it would be the same as using PHP or similar, where you actually have to spend more in developers.

both go and python 3 arent good solutions while Node 2 seems to designed for specific tasks

cant wait

>>926595 (OP)

He looks like a cokehead these days.

LOL

>>926655

Simple. Make your web server check if the ip that's downloading is from a datacenter. If so send the vulnerable one, else send the clean version.

>>926681

Is there any reason in particular why that's more of a problem in that situation than in the usual situation where you install packages on deployment?

>>926681

I feel that in practice you would just import stuff from CDNs.

Just like you don't just copy a link from wherever and paste it on your HTML, you usually just import stuff from CDNs.

>>926685

Because package are apparently not signed

Using node ever since 0.10.

Even IF deno comes to fruition, it won't be able to get any traction.

Node is too big and too liked to be able to be replaced by something only marginally better. It will all boil down to how much companies would be able to save by migrating from node and how much that would cost. As it stands, it just wouldn't pay off.

If we pretend Node doesn't exist, it'll go away.

>>926646

>python

Cancer.

Either POSIX scripting or a programming language.

>>926766

>python isnt a programming language

>>926769

Last time I heard python is scripting.

>>926766

POSIX is insane and barely suitable for scripting unless you extend it.

Scripting languages can be programming languages.

>>926772

>Last time I heard python is scripting.

You are an imbecile.

>>926784

why is the tech field filled with these zealots that must argue basic facts? it's like arguing with a sjw or worse.

>>926787

It's reasonable not to call sh a programming language, because it has severe limitations that make complex computations in it a pain.

It's reasonable to call Python a programming language, because it lets you express any computation you could express in a language like C without much more fuss (and usually with much less). You can write full programs with it, they just won't be very efficient programs.

>>926793

That doesn't mean it's not a scripting language.

>>926756

I guess it will be like this. Inertia is just too strong in the IT world. It takes too much time, resources and effort to change some system already in place and being used by people who got used to it. If there's no huge benefits to change, companies will just shrug and keep what they already have.

>>926829

>Inertia is just too strong in the IT world

That's why GCC makes optimization assumptions for CPUs from 2000 when making generic 64 bit x86 code.

They should up that a bit, 2004 would be a nice starting point but nobody is using those weird as fuck 64 bit steppings of 32 bit processors anymore. Make assumptions for 2009~ CPUs instead and you'll benefit both 10 year old systems and current systems without touching early 64 bit CPUs too much.

>>926646

What's wrong with TypeScript?

>>926672

>>926766

Or Kotlin if you don't like Python 3

>>926933

Too JS-like

>>926766

>POSIX scripting

literally worse than even js

>a programming language

Python 3 is one of them

>>926933

Many flaws of the JS are still there for compatibility reasons.

>>926793

>they just won't be very efficient programs

The efficiency should be measured on real workload, and if the slowdown is measured in fractions of a percent, then it's obviously not worth dropping the benefit of using "easier" language.

For many tasks they will be actually efficient.

And even for some number crunching tasks: a typical example is https://github.com/magicgoose/simple_dr_meter/issues/4

Not sure if ffmpeg guys catched up yet (I need to ask the author to test again or do it myself), but a piece of python code which calculates the same thing was consistently faster when compared to ffmpeg, which is written in C and was compiled with full optimizations. It would not be possible if what you said was true, right?

>>926820

Because >>926766 said Python is neither POSIX scripting nor a programming language, it was saying Python wasn't a programming language. The rest of the subthread was about that claim.

>>926980

I meant efficiency in the simplest sense - efficient use of memory and CPU time. I agree that that's not a useful metric to decide whether the language is appropriate to use for a task.

Your example is cheating by using numpy. There's nothing wrong with that, of course - I love numpy - but it doesn't tell you much about the speed of Python because it's really just Python wrapped around C, C++ and FORTRAN.

A better example would be an I/O-bound program. youtube-dl isn't going to get much faster whatever language you write it in, because it spends most of its time waiting for data to arrive.

Most code I write is Python.

>>926978

Kotlin, Swift, Go, Flutter, Java, C# or Objective C.

Or on a more serious note, Python 3/4 or Perl 5/6.

>>926668

Underrated post. Fucking money worshipping jews screw us once again

>>926675

Nobody actually believes that soibois write code.

>>926668

Why not purescript then anon? It works on the front and backend.

>>927314

At a glance, it looks like it's easier to teach Purescript to people who already know Haskell than to people who already know Javascript.

>>927318

But why would you ever want javascript developers?

>>927349

Same reason people want C programmers.

As awful as the languages are, they get closest to the way the runtime Really Works™.

>>935500

w.r.t. application development I'm thinking of Electron apps.

But I've seen criticism of it for server side stuff too.

>>926646

snekfags btfo

>>927314

Actually this.

>>927349

And this.

>let's use a language with no metaprogramming and no primitive integer type for anything beyond simple one-page scripts

>>935603

Javascript (ES6) actually has metaprogramming.

oh great more webdev garbage

if I am made King by Divine Right I will execute every single last one of these greasy javascript niggers. javascript is satanic cancerous niggerware

>>935500

The best invention since orgasm, tbh.

>>935500

Node really isn't that bad. It's the webdevs that think that they are real devs that is the disaster.

>still no threads

>computations still clog the fuck out of the event loop

Yeah no thanks.

>>935744

This. Deno is Node with slightly less HIV but it's still HIV+ so who gives a fuck.

import { test } from "https://unpkg.com/deno_testing@0.0.5/testing.ts"

Specifying the version at every import will make dependency updates really painful, but I can understand wanting to get rid of package.json (and npm)

>>926595 (OP)

Yeah well I'm still cumming over C++20

>>935666

Well it doesn't have macros and it still doesn't have an integral data type separate from "Number"

>>936255

why?

>>926933

TypeScript and JavaScript are like C++ and C.

>>936805

>Plan 9 the network operating system

>something good

Every Plan 9 larper ended up either being dead or working at Google, not just Pike. I'm not a Lisp faggot, but I'm fucking 100% sure the whole Unix paradigm spawned this SaaS and cloud "service" kikery because all these people had a literal baby duck syndrome and refused to develop personal computer software because they were used to using mainframe from a dumb terminal, be it AT&T UNIX, Bell Labs Plan 9, or Google mainframe. Consider it an extension of same oppressive sociopathic conformist Silicon Valley cult. Microsoft, blame them for thousand years, partially escaped this cancer in those years because they constantly moved cities before settling in Seattle and Redmond.

Also the quotes are from this site for those who are interested:

https://usesthis.com/

It's full of "coffee shop Macintosh transvestite WWW developers" as youngsters would say, but has a nice batch of free software and hardware developers's interviews including Stallman and Torvalds.

Here is an example of your average Intel engineer: https://usesthis.com/interviews/bo.adepoju/ :^)

>>936838

Sauce on the wallpaper?

>>936805

Kek, saved.hope he has GRIDS too

>>936805

>quotes on the right

He's right. There's no reason our networks are not distributed and reliable without trust.

I think it's absurd that the left side boasts about Plan9 but uses those quotes as negatives. As if distributed computing wasen't the entire point of Plan9. You should have the power of big irons on your phone, there's no reason you shouldn't. The device itself need not be that strong, you just need to be able to cooperate with one that is and that's 100% possible to setup and use practically, since the fucking 90s but we still insist on building 1 machine per person, while not utilizing them effectively.

When you really consider how much computing power is out there and how much is used, it's ludicrous.

>>936826

Say what you will about Plan9 itself, in concept it's good. And that's all you can ask from a research OS. Even if we're not using Plan9, we should be using something with the same ideals. I can't see anything wrong with remotely aggregating resources and abstracting away the concern of if it's actually local or not. If I have access to resource X, I should have access to it on every machine I interact with, without having to manage the state myself, the system should take care of all of this.

>but cloud computing is bad!

Maybe when companies run 1 giant cloud and ask you to use it. But this is not that, this is you setting up or joining multiple clouds, and making them interoperate. Maybe you host an entire machine, maybe your friend John lends you his processor, maybe you pay Google for access to a disk pool. In any case, it shouldn't concern you more than the initial setup. You should only have to worry about how to authenticate with the network. A network that doesn't have to be inherently centralized and shitty as it is now. A network where losing connection to said networked resources, is unlikely.

>>936754

I cum pretty easily when there's updates

>>936595

Javascript (ES6) has Macros and ES2017 has int32, int64, float and all the standard numerical types.

>>936942

>The device itself need not be that strong, you just need to be able to cooperate with one that is and that's 100% possible to setup and use practically

This is the reason why I use a laptop as a personal terminal device for input and display. I have a personal Lignux server that's intended to carry out batch computations.

>>937913

It has neither of those things, you liar. Sweet.js can be used for macros, but that's not part of javascript. Typed arrays allow you to store arrays of integral or floating-point types of various sizes, but there is no primitive integral data type. Rather, there is just "Number."

>>926611

Pike's about a week from cutting his penis off for progress. Hail Theo.

>>938582

Theo is a swap german and a cloghumping virgin cuck rat

>>926595 (OP)

It took node almost a decade to be not-shit

>>936805

The cuckening

>>936805

JUST

Is he planning of deprecating npm support? b/c that SJW-infested hive of horseshit needs to expire.