[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / siberia / hobby / tech / edu / games / anime / music / draw / AKM ] [ meta / roulette ] [ cytube / git ] [ GET / ref / marx / booru / zine ]

/tech/ - Technology

"Technology reveals the active relation of man to nature" - Karl Marx
Name
Email
Subject
Comment
Flag
File
Embed
Password (For file deletion.)

Join our Matrix Chat <=> IRC: #leftypol on Rizon

| Catalog | Home
|

File: 1632339083981.png (227.08 KB, 1008x736, filepicker.png)

 No.11670[Reply]

Greetings, chinlets and incels. I am a man of the name Dimitrioyivch Khaladrenis. And I have been greatly frustrated with the filepicker situation on linux as of late. Kdialog doesn't force itself on most applications save those which support it, (Only firefox) and the gtk filepicker has no thumbnails. Kdialog also has the problem of being tied to KDE so it's not very lightweight, not a sucklessfag but sometimes one gets the urge to activate ye 'ol i3 but still wants to have a file dialog with thumbnails, if you catch my drift.

Anyways, my solution to the File Picker Question is to basically code one myself. However I am also a NOCODER and don't know where to start or how GtkFileChooser and Kdialog actually work, nor do I know the process of creating a xorg window from scratch, or how necessary Qt or GTK are in that process. The finished product should be able to:
- Have thumbnails
- Let you type in the path of the file you wish to use
- Have shortcuts to the folders where you keep your photos/webms/etc
- Open instead of GtkFileChooser on all programs
- A search function for the current active directory
7 posts omitted. Click reply to view.

 No.11803

>>11670
Yoshii from texhnolyze?

 No.11806

>>11800
Been thinking of making it proprietary due to the fact that
>>11722 doesn't seem to think I can do it. So it's going to cost 1 ETH to buy the binary and even then I'm going to put it under a restrictive EULA.

 No.11814

>>11722
>Why didn't Ubuntu devs fix the file picker? Because even they don't wanna do it.
Sure, but why they don't want to? Not because it's impossible but because it isn't the job of package maintainers to develop software. I don't doubt many hobbyist patches out there are garbage. But what is a file chooser anyway? It's basically a dialog version of a file browser, right? So if so many file browsers on Linux support icon view, why would it be so hard to just port that functionality to file chooser? Make them share the same thumbnail cache and settings for which directory gets displayed in the icon view. The solution is trivial, but it does require a lot of work. Not mentally challenging work but the mundane work of porting, rewriting, refactoring, testing. And for what purpose, what guarantees that their fork will have any effect on upstream or distros? Why don't the Mint devs do it? Because it would require maintaining A LOT of packages which they currently very happily delegate to the Ubuntu maintainers and for good reason. People already have enough on their plate maintaining the existing forks and patches of GNOME fuck ups, considering multiple whole new DEs and programs appeared because of GNOME devs.

 No.11818

>>11814
The real reason is that Red Hat plans on creating a proprietary filepicker for GNOME, they will call it FILECHOOSIX and rent it out for a $1000 monthly subscription to enterprises. They have played us for fools. The only way to fight back is to abandon GTK or fork a truly libre version without corporate meddling. That's our only choice if we want to save the future of filepickers on the linux desktop.

 No.11840

File: 1632836390195.png (2.17 MB, 875x1196, ClipboardImage.png)

>>11818
I pray to st. ignucious they would actually do that. Let one thousand forks bloom.



File: 1632644242504.png (178.56 KB, 940x198, header1.png)

 No.11813[Reply]

https://arcan-fe.com/
>Arcan is a single-user, user-facing, networked overlay operating system.
based or cringe?

 No.11816

>>11813
smell like a fed honeypot

 No.11817

Extremely based, it goes way beyond current isolation-based solutions (like Qubes) which tackle only one aspect of the problem. But if I understand it correctly you'd have to rewrite huge chunks of the user space to fully benefit from this design.

This is a good read as well, has some more interesting ideas:
https://www.divergent-desktop.org/blog/2020/08/10/principles-overview/

>>11816
LMAO, how?

 No.11827

>>11813
Super based and the best hope the Linux desktop has post-Xorg, IMO

 No.11828

>>11817
>But if I understand it correctly you'd have to rewrite huge chunks of the user space to fully benefit from this design.
It has Xorg and Wayland compatibility layers, though maybe not all the "Arcan features" are available with those



File: 1624038678945.png (17.74 KB, 579x382, 34567899876.png)

 No.9418[Reply]

Hello, tech, I'm looking to develop a leftypol textboard for the gemini protocol. I am curious about where I should get started, or, what language would be optimal for this. I also don't really understand how to get one language to interact with another language. (Gem text and Perl for example) I don't really want to do anything huge or something that will blow up I am just a hobbyist looking to do something fun and interesting. thoughts?
7 posts and 1 image reply omitted. Click reply to view.

 No.9434

>>9433
Based.

Here's another textboard script that tries to remain as close as possible to 2channel but with a database and QoL features.
https://git.bienvenidoainternet.org/bai/weabot/

 No.9709

>>9429
>That site looks like shi and is just one giant wall of topics and replies. The formatting is horrible.
Gemini markup doesn't give you many options for formatting, just some headings lists and links, and clients can render them all differently.
You're missing the point if your criticism of it is "it looks fugly".

 No.9711

>>9709
I want something more structured like station.martinrue.org

 No.9716

>>9418
>he was a windows user all along

oh no no no no no

 No.11822

If anyone is still interested in this. I found a gemini textboard written in rust.
shitty, immature license though
https://git.kiwifarms.net/CrunkLord420/gemboard



File: 1630964531349.png (43.78 KB, 800x640, 800px-X.Org_Logo.svg.png)

 No.11103[Reply]

How do you laugh at win users for security when we use shit like pic related?
17 posts and 1 image reply omitted. Click reply to view.

 No.11134

>>11131
werksforme

use the cfs scheduler and ionice.

 No.11135

>>11134
*cfq scheduler

 No.11136

>>11135
ah, it seems they replaced it with bfq.

 No.11801

>>11133
You can't make ANYTHING work on modern hardware (as in support all the video, wifi cards and shit) unless you are a corporation
It would be a nice undertaking imho, a piece of code that's just completely illegal but benevolent and usable. It would warm my heart just knowing there are people rebellious enough to do this
>>11134
And on windows you don't need fine tuning for that

 No.11804

>>11103
I bareback the interwebz running wind0ws



File: 1631893826186.png (94.85 KB, 465x421, pspherexmpp.png)

 No.11598[Reply]

There's a new Jabber/XMPP server in town, running on Prosody (https://prosody.im).

prolesphere.one | d7dchu2gnb4xce7y7vwmulonr4k2wh2dmobqyitwwjeiojfotrvw37yd.onion

Registrations are open, and so is room creation. The MUC (Multi-User Chat) server is at hive.prolesphere.one

The server supports OMEMO and OTR encryption, or should. Expect some growing pains in the early stages, there might be kinks to iron out so server may restart spontaneously. You can report any problems here in the thread or by messaging me at [email protected] (not an email).

You can find a list of XMPP clients here: https://xmpp.org/software/clients/

I use Gajim on desktop and Conversations (available from F-Droid) on mobile.
5 posts omitted. Click reply to view.

 No.11656

>>11655
poezio and profanity according to my local xmpp guru. irssi-xmpp is also an option

 No.11657

Won't this just disappear one day like the board?

 No.11659

>>11657
That is a fair criticism. I did let the hosting expire. It was a super busy period and it slipped my mind. To that end I am going to keep a better eye on the deadline and fund the crypto wallet + pay for the next quarter way ahead of the deadline so it doesn't happen again. I have also set a reminder on my phone.

 No.11662

>>11655
Pidgin or gajim for desktop, conversations on mobile

 No.11728

>>11657
you can always register a new account on another server and re-add everyone in your roster



 No.11692[Reply]

I'm less interested in the biology of the deep sea(though it does help with understanding the origins of life alot) but rather the physical phenomena that take place at seriously deep depths, what strides should society take as a whole to venture into earths true depth and see what lurks not just below the surface but past the point where water pressure elevates to such levels where one could die simply by being in the right position either from pressure or from earths gravity and what kind of things can man find there

 No.11702

Youd die if you tried breaking through the sea floor to see what lurks below



File: 1624167488802.jpeg (7.54 KB, 293x172, hands.jpeg)

 No.9462[Reply]

Why does Rust get hate here? I know its a memed language but is there any good reason not to use it other than be contrarian?
32 posts omitted. Click reply to view.

 No.11058

>>10949
I don't think the language used in the implementation of the compiler has ANY impact on the language it compiles. The language design (usually) is done before even a lexer is written

 No.11073

>>11058
It does influence how the language is compiled, its internal data structures and its function primitives. But more importantly if, we return to the example of fasm, it would only need to interact with code compiled from C during runtime, yet the lowest level of high level programming would still gravitate towards C-like patterns, because widespread modern hardware is for all intends and purposes "built to run C", meaning imperative code with variables of wordsize or below and continuos memory access, thanks to the bottleneck of read/write operations through a memory bus. http://www.loper-os.org/?p=46
Optimizing for hardware reifies its bias towards this programming pattern, even more so with POSIX compatible software. I programmed in C, so I cannot unsee.

 No.11667

Rust gets hate because a lot of its userbase is basically zoomer train stations, furries, etc hyperliberal types that'll remind you to say "person" instead of man or woman. But it's a better C++ otherwise.

 No.11668

>>11073
The processor world moved on from C, now compilers do black magic to turn "imperative code with variables of wordsize or below" into SIMD code. If you want to write it by hand, all the architectures have nonstandard extensions to C to enable vectorization, but they are very clumsy and very obviously don't fit into C's model. A language that lets you easily write portable vector code could become really big.

 No.11674

File: 1632346846523.png (18.04 KB, 429x470, simd_cpu.png)

>>11668
How does SIMD fundamentally change the dominant programming paradigms of C?
As far as I see, it simply amounts to loading data into the SIMD registers, potentially at a size different from wordsize, which could be implemented by loop unrolling of continuous memory access in the form of duff's device.
You may consider how high of a performance gain multimedia processing, the main motivation for adding SIMD extensions to x86, receives from caching, which indicates its proximity to the continuous memory access pattern.
>The high cache hit rates of multimedia applications can be contributed to the following three factors. Most multimedia applications apply block partitioning algorithms to the input data, and work on small blocks of data that easily fit into the cache. Secondly, within these blocks, there is significant data reuse as well as spatial locality. The third reason is that a large number of references generated by multimedia applications are to their internal data structures, which are relatively small and can also easily fit into reasonably-sized caches.
https://www.researchgate.net/publication/221596588_A_study_of_memory_system_performance_of_multimedia_applications



File: 1632335889834.png (187.46 KB, 410x499, thoughtful.png)

 No.11704[Reply]

Why is it so hard for chans to be profitable and make money compared to other kinds of "social media" ? is it just the edginess? Or lack of data mining that comes with anonymity?
11 posts omitted. Click reply to view.

 No.11716

>>11713
Unbased

 No.11717

First forget existing forms like image boards and social media exist.
Now think about this starting from profitability, and work backwards.
You want to maximize the rent you can extract from advertisers.

1. Must be image or video based, so that ads can blend in with the other visual content.
2. Norms to funnel people to create content most similar to ads. Videos should be around the same length as an ad, so everything blends together and can't be re-separated.
3. Must be algorithmically moderated so you don't need to pay any workers.
4. Must have strong identities for two reasons.
- Minor reason: Increase the quality of targeting metrics
- Major reason: So brands can go down and mingle with the plebs. Force people to interact with brands using the same tools they interact with people (Likes and retweets).
5. Captive audience. You want people who have already had their autonomy constrained by circumstance, in this case platform. (Tech illiterates use phones and can't install adblockers to fight back).
6. Platform level data harvesting. Adverting companies want to see reports on where their money went, even if it's 100% a bullshit fantasy (Which it is). The more plausible deniability embedded in the system, the longer the system will operate without tearing itself apart.

Well done, you've invented instagram or tiktok. And not 4chan. Even reddit is clearly showing its age, and fails to be the most modern and extractive social media it could be.

 No.11718

>>11717
1. Chans require an OP which have images or videos (usually)
2. It's totally possible in theory, though not usually done, to have threads/posts which are the equivalent of "sponsored tweets" aka ads which look like normal tweets.
3. Even FB/Insta have manual moderation to some degree. Paying workers, esp. thirdoids to delete spam is way cheaper than making muh deep learning facial recognition spam deleter (tm) latte sipping silicon valley engineers that you have to pay $300k a year.
4. you're right, chans will never have this. It's inimical to the nature of chans. Even with accounts there's limited knowledge of the actual person. Pseudonymous platforms like twitter/reddit are much less profitable than insta/fb which have data mining up the wazoo.
5. Phonefagging is actively discouraged on chans by lack of official apps and also bad mobile CSS. Android has 4chanx at least. This could be corrected though by having mobile friendly CSS.
6. see 5

Your argument basically boils down to platforms which collect the most info and data mine the hardest are the most profitable

 No.11719

>>11718
Didn't read motherfucker. That wasn't an argument that needed a autistic point-for-point refutation.

 No.11720

>>11719
ok whatever the point is with the ability to host imageboards cheaply you would think they'd manage to make it work. I read somewhere hosting 4chan costs $20k a month. This is BS. for 4 app servers? Even with the beefiest servers and a load balancer any retard can create a chan for like a few hundred bucks a month MAX. Even bunkerchan at its peak only cost like $120 a month. With shit like that you'd THINK they'd be able to make SOME money on it. At most it should be like 1000 or 2000 dollars a month



 No.11144[Reply]

Why are 90% of the threads here dedicated to programming? Code has hardly changed since the early 1900s with most of the actual development in computer tech being in processing power and multipurpose functions but computers like that are use by the government for research in medical shit, heavy machinery(rockets, military and exploration tech, nuclear equipment etc that kinda shit) and particle physics than whatever the fuck most people on here are posting about. What's the appeal in the never ending barrage of different languages and shitty undergrad textbooks?
20 posts omitted. Click reply to view.

 No.11604

>>11575
Parcimonie-like[1] RSS feeder. Which means each feed is fetched at a random interval through a throwaway Tor circuit, the goal is defense against profiling you on the basis of your feed selection. Its requests have the exact same fingerprint as Tor Browser at the safest security level. Also supports custom plugins so that you can convert whatever page you want into a RSS feed in case the page doesn't provide its own feed.

(But you still need a separate RSS client to view the fetched feeds locally, like newsboat[2] for example.)

[1] https://github.com/EtiennePerot/parcimonie.sh/blob/master/parcimonie-original-design.md
[2] https://newsboat.org/

 No.11620

>>11563
Nobody stopped them from contributing like programmers contribute.
But UI and design types of people think of themselves as "artists" and are heavily invested into the whole "you can't copy or share without permission!!!!" spook and their philosophy is at odds with the shared nature of free software.

 No.11622

>>11575
I have a few projects open sourced

- A Rust EPP library for domain registrars
- A browser for git repositories on Gemini written in Rust. It allows you to browse trees but it needs a commit browser to be added.
- A Capistrano-like deployment automation project that's written in shell script so there are minimal dependencies. It allows you to deploy to docker hosts, registries and kubernetes clusters also. I like Capistrano but I didn't want my non-Ruby projects to have a Ruby dependency. It's been working well in production use doing 30+ deployments a day without issue.

I won't link the repos here because prefer to maintain (pseudo?) anonymity, but you can easily find them since there aren't really many EPP libraries for Rust available apart from mine.

 No.11634

>>11604
thats really cool

 No.11663

>>11559
UX and UI designers are not necessarily "normal" people either. Just look at some recent trends in the last decade, it's aesthetics over usability, in many places shit's getting worse and worse.



File: 1630847512359.jpg (121.49 KB, 1280x720, slackware.jpg)

 No.11054[Reply]

11 posts omitted. Click reply to view.

 No.11095

>>11082
Gentoo is in no way historically significant.
Slackware on the other hand IS.

 No.11096

>>11077
By switching from Arch to Slackware. Slackware is better than Arch in every sense of the word. It's easier to install, more stable and the 3rd party packages (slackbuilds) are much more safer and trustworthy than the AUR will ever be.

 No.11142

Not including stuff related to any desktop environment like GNOME, is it true that distros like Arch, Slackware and Gentoo are any more performant or easier to maintain than Debian, Ubuntu or Fedora?

 No.11529

>>11142
No
and they're harder to maintain
except maybe arch which is on par with debian etc

 No.11661

>>11142
Nah for gentoo for sure. I mean, gentoo is in some aspects slower than Arch and others. For example, PGO flag was borked for firefox for many years (I think it works now but idk) and all other distros like Arch have Firefox with PGO compiled and that's 6-10% performance improvement according to mozilla. There are some other packages that also have borked PGO right this second (and compiled with PGO in all binary distros) so basically yeah, gentoo is a meme. Or that is to say, maybe they should stop supporting 100 architectures and focus on relevant ones.



Delete Post [ ]
[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / siberia / hobby / tech / edu / games / anime / music / draw / AKM ] [ meta / roulette ] [ cytube / git ] [ GET / ref / marx / booru / zine ]
[ 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 10 / 11 / 12 / 13 / 14 / 15 / 16 / 17 / 18 / 19 / 20 / 21 / 22 / 23 / 24 / 25 / 26 / 27 / 28 / 29 / 30 / 31 / 32 / 33 / 34 / 35 / 36 ]
| Catalog | Home