/g/ - Technology

A channel with a mix of documentaries: https://tube.privacytools.io/accounts/documentaries@vidcommons.org/video-channels

>>316

Bumping for interest, since I never found anything worth while myself. I do appreciate the service however. The additions in this thread so far seem interesting.

Looks like the people at Framasoft realized how hard it is to find anything on PeerTube and made a semi-curated search engine for it: https://sepiasearch.org/
Announcement: https://framablog.org/2020/09/22/sepia-search-our-search-engine-to-promote-peertube/

This is some netlabel's PeerTube, they upload music videos to it: https://v.basspistol.org/videos/local

This site has "edutainment" videos, if I understand correctly they are uploaded by youtubers (i.e., it's not an unofficial mirror).
https://tilvids.com/videos/watch/69da6d3a-b56b-4b22-800d-8eb1530f4394

Sounds like a cult

>>421
They adhere to a certain sense of humor as far as I can tell.
>>422
The CIA does use it.

>>428
"cool" cult

>9 days ago
Give us an update, OP. Are you on your way to becoming the next Stanley Lieber?

It sounds like you could use Qubes OS, Whonix, Tails, or some combination thereof.

Why not install WINE or even use Steam Proton?

Why can't you resize the partition?

Resize your Windows partition and encrypt your linux partition with luks or something like that. Instead of default windows bootloader get grub ot refi

shrink your windows partition and leave some space for linux. if it doesnt let you, it might be too fragmented

I rarely have any issue with youtube through tor.

They're going after youtube-dl.

This is a pretty exhaustive place for comparing VPS providers: https://lowendbox.com/

Also, I just found this dedicated server provider: https://www.kimsufi.com/en/servers.xml

The prices are insanely low considering that these are dedicated servers, not VPS's. For example, you can get 4GB of RAM, 2TB disk space and a two-core Intel CPU for €6 a month.

>>430
cope

I'm glad u asked.

>why don't software engineers just write proofs instead of programs
Well, technically, they do write proofs, since programs are the same thing.
The things is, they usually prove something like:
"given integers x,y, either floor(x/y) is an integer or RuntimeException".

I guess are asking why their proofs almost always prove this extra " or RuntimeException" (and actually there are usually more such terms).
The reason is just that it's easier to not be as precise, at least with the current technology.
For example, if i can use (int)( x*(x+2)/2 ) without having to prove that x*(x+2)/2 is an integer.

There are some languages whose programs really are proofs (e.g. Coq),
but they are more difficult to read and write, have less access to libraries, etc.

I think there should be a way to write safe, yet readable and efficient code,
Post too long. Click here to view the full text.

>>389
That's not entirely correct. What you describe is the typechecking, which is equivalent to the correctness check of a proof. But the statement that the proof proves is not this, but the type itself. If your proof is "floor(x/y)", what it proves is that if you can construct two integers, then you can construct an "integer or RuntimeException".

It's easier if we put it in a method. The following three methods have the same type and therefore prove the same statement:

int nonsense(int x, int y) throws RuntimeException {
    return floor(x/y);
}
int nonsense2(int x, int y) throws RuntimeException {
    return 0;
}
int nonsense3(int x, int y) throws RuntimeException {
    throw new RuntimeException();
}

Post too long. Click here to view the full text.

>>390

>That's not entirely correct.
What i should've said is that, in hypothetical a program,
the expression floor(x/y) proves that, given integers x,y, either you can make an integer or throw a RuntimeException.

>throws RuntimeException
In every Java declaration,
there is an implicit RuntimeException,
since they can be thrown anywhere without declaration,
so any method's return type should be considered a union of the stated type with a RunTimeException throw.

> return bullshit(x);
I guess we should add an "infinite execution" to all our RuntimeException terms.
You could also return null,
Post too long. Click here to view the full text.

>>368
>>369
complacency kills, but at the same time not everything has to be a labor of love. if i'm not getting premium pay i don't generally put out a premium product.
if it's my own project it's a different story though

>>368
Do you have a linter hooked up with whatever your team uses for code reviews? If not, do everything you can to get one, it makes everyone's life much easier. But yeah, that sounds miserable.

>>371
Yeah no I agree with you. I think I perceive it this way since I'm only at the starting phase of my career and lack the exposure. I know that most pragmatic approach is to save my autism for my personal programming projects, rather than being a nuisance to everyone at work.

>>372
We do actually, but it mostly cleans up formatting errors and gives some warnings about unused variables and such. Not really that helpful. I mostly have a problem with their portrayed disinterest for structure and cleanliness. I perceive their mentality to be something along the lines of: "My first attempt did what it was supposed to, so why take a moment to consider it and possibly iterate it one more time.". That being said I realize that you have to draw the line somewhere. You probably get my point.

That's enough about my complaints. Reiterating OP's question: What do you look for in reviews?

>>366
In my last job, I was as picky as you, OP. But when it came to fixing all the findings, we were told to only fix potential bugs, and leave all the violations of the coding style guide in place. I especially hated it, when people write code, that doesn't even fit on a full HD screen with a small font. Like what were they thinking? I would be ashamed to deliver such code. The real fun begins when you try to do diffs side by side with such a mess.

I also felt like i was too picky all the time, because nobody else seemed to care, but on the other hand, this software was low-level code for devices peoples lifes depend on.

You probably don't want to know what shitty code is inside health devices like pacemakers or insuline pumps and what not.

The X Window system had a lot of problems but most are solved now really, beyond being sort of complex and "insecure" (doesn't really matter, since Unix-likes are about as insecure anyway). Wayland is supposed to break things down into some D-Bus/XDG/logind nightmare or something, not entirely clear to me, maybe someone knows better here.
Ignoring all other options and going directly for GTK and Qt (Qt isn't nearly as bad as GTK though). GTK and Qt go through major version changes and impose change on programs which creates instability and extra developer work. GNOME has been changing its design paradigm and everyone else has to pick up their turds and deal with it.
GIMP is not even bad other than the fact that it uses GTK. Single window mode, dark mode, GEGL options, Mypaint brushes, its all been default for quite some time now, but Adobe users just won't switch away from an 8 year old program (the pre-cloud version of Photoshop). Never used Photoshop, the most complex proprietary graphics editor I have used is Paint.NET, but GIMP is more capable than that one. Krita is a little easier for painting and drawing but it has a worse text tool last I checked. I remember people whining about CMYK for years, but its in Krita and GIMP now as far as I know (there was a GIMP plugin for CMYK for years). I think they just don't want to learn a new tool, there will always be excuses, but they don't really mean much since these people don't plan to switch ever.

>>363
cont'd
I like Eaglemode, graphical Emacs (the X kind without GTK), and some of the X window managers as well as mpv. Links (browser) has a graphical mode but it looks much like the curses interface. Worker is an interesting file manager and a good example of a classic X11 program that never died out, but its a bit too orthodox, so I don't use it.
Once you get away from the big GUI silos trying to replicate some other system semi-successfully, you can see the advantages and disadvantages of the Unix-like approach. Unix-likes don't really care much about graphics so there is a lot of complexity in the whole pile of display technologies. Ideally they're just trying to display a TTY, not a graphical environment with GL acceleration and programs capable of displaying OpenGL or Vulkan stuff with specialized hardware and drivers. It took many years for Mesa/drm (not digital rights management) in the kernel/X to work this well for a lot of people, and there are still problems. Desktop software wise, people had to write a lot of standards, XDG/ICCCM/EMWH and so on, and they're not really seamless or without problems, so its kind of terrible for an Apple, Windows, or BeOS clone. But on Unix-likes, you can have it your way usually, you can have your own window manager of choice that works well for you and can be modified to fit your needs, rice Emacs or Vim or whatever, tile those windows, etc.
Oh and NVIDIA doesn't care much about Linux or other open source platforms so CUDA isn't really as relevant (requires proprietary drivers), and OpenCL usage is limited, so there are less programs using or related to that sort of thing.

>>364
>Eaglemode
Good taste.

>>256
I went through something similar XD

>be me, 9 years old
>get a cheap Dell laptop for Christmass
>AwYeah.jpg
>Immediately search up "Minecraft for free"
>go to the first website, already looking sus
>download it without a second thought
>computer becomes very slow for some reason
>see that I have a new app with a pink dolphin logo
>coolbutididn'task.png
>didn't get minecraft for free but I got a lot of popups of sexy gurls on every webpage ;)

We all learn about internet safety eventually…

>>37
>What is the worst thing ever happened on your PC?
Me.

was trying to install drivers for my disc drive on debian and I fucked up my xfce and got it to boot into the command line by default. I had to type “startx” to get my gui to appear again.

I think my PCI slot is close to dying, my graphics card only works half the time. I'm too poor to buy a new motherboard, let alone PC…

Freaky. I think they draw a couple images and use animation to fill in the rest

>>308
AI software. Similar to what's going on with the deepfakes.

>>358
This just looks like Live2D to me, same sort of thing as what all those VTubers use. This type of thing doesn't require anywhere near the complexity of deepfakes.

That's what I use, so I guess I win too?

Using Ryzen AMD CPU as well here.

Apple announced that they will be using their own chips; I wonder how that will effect x86's monopoly on desktop computing.

>>319
It probably won't. Macs already had compatibility issues with things made for PCs (e.g. Linux), so I doubt widening that gap further will change much. Then again, I tend to be continuously surprised at what kind of cues tech companies take from Apple. Like, it didn't surprise me at all when iPhone abandoned headphone jacks, but I was absolutely baffled when other phone companies followed suit.
If anything breaks the x86 monopoly, I really hope it's something open source like RISC-V and not just companies trying to copy Apple's dumb bullshit. Who am I kidding though, that's not gonna be the case.