@fluffy@BrodieOnLinux@yes Debian is literally the "compromising eternally" version of Linux IMO. The whole reason comes down to the fact that you can add the non-free blobs if you want to, as opposed to the FSF endorsed distros which are 100% deblobbed (since in the eyes of the FSF, blobs are only "okay" if it's "embedded" into the hardware).
@PhenomX6@BrodieOnLinux@fluffy@yes I used to run Debian servers for decades, nowadays my servers are a mix of Devuan and OpenBSD, the only reason is because it's impossible to pick up Onion and I2P hostnames, and move them to some place else.
@PhenomX6@BrodieOnLinux@fluffy@yes Sorry, I got interrupted, and forgot to say what I wanted to say. Debian has made so much fatal mistakes, packages are forever heavily outdated, it's SoystemD, and it's the base for some of the worst distro's ever made (Ubuntu etc).
@charliebrownau@PhenomX6@BrodieOnLinux@fluffy@yes IPFS requires you to dox your IP location, and requires new hardware to prevent it from crashing (and prevent you from using that hardware long term, because planned obsolecense). As for Zeronet, I heard it's proprietary (though I could be wrong). And most of the TLDs managed by OpenNIC aren't registerable (and those that are just don't work).
I2P is good.
What we need is have 1 simple self hosted page on the clearnet that instructs you how to access your actual websites (changing DNS or hostfiles, installing something like Tor or I2P etc).
It looks like the future of the internet will be hidden and invite-only, so only those with a brain can actually find it, like how it used to be.
@ryo@PhenomX6@charliebrownau
Last time I peeked into the OpenNIC community (probably 7+ years ago) I remember the situation being where everyone except like one or two TLD operators all using a very outdated, broken, and unsupported PHP application (I’m talking like “register_globals”-era tier) that was written for OpenNIC use by someone that had since moved on. There was periodic interest in someone to come along and do a full rewrite of it, usually never reaching any fruition, meanwhile one of the lead people having a total and absolute deathgrip that any rewrite absolutely just HAD to target PowerDNS (I believe) as a backend authoritative DNS server, all because PowerDNS had the feature of querying a MySQL server behind it for any DNS records… ‘therefore clearly the most superior option’ since you could just set DNS records in a MySQL database instead of having to ‘deal’ with zone files…
Meanwhile BIND has had the capacity of dynamically creating and updating new DNS zones in a running server since ~”9.7.2-P2” released (around late 2010?), without having to reload all zone files, but nope, that’s not good enough, it “HAS TO be PowerDNS!” Either way, I’m sure someone else could come along and do better. The other fun thing with DNS is that there’s other serializations now, like DNS-over-HTTPS (JSON), where you could even have a braindead web developer be able to write some unusual alternative DNS network (versus just a typical alt-root DNS network using standard conventional DNS), and also not have to deal with any public DNS servers being weaponized for reflected DoS attacks per the inherent issue of DNS, UDP, and the source IP in a IPv4/IPv6 being forgeable for any networks that don’t conform to BCP 38.
I agree on commending I2P, and I do also commend Tor as an option, but would weigh I2P over Tor.
And yes, I agree on the prediction of the future outlook of the internet being a scatter of separate invite-only overlay networks, using the internet merely as a transport.
@charliebrownau@PhenomX6@BrodieOnLinux@fluffy@yes In that sense it would make more sense to have no internet at all then, as you can then physically contact people in your local area. But it does make sense if you consider that we're living in an endless time loop where first there's no technology at all, then technology becomes more and more advanced over time, then it gets turned from a tool of communication to a tool of mass surveillance, then AI comes in, then the world goes "fuck technology!", and we're back without technology again, only to then crave for a way for John Smith to communicate with Tanaka Yamada, Luigi Spaghetti, and Dmitri Kurvablyat all living in places far away from each other.
@charliebrownau@BrodieOnLinux I think PeerTube is the only true way forward to be honest. As soon as JoshwhoTV becomes big enough, they might probably bend the knee too at some point. Not saying because I don't like the platform, I'm saying that because it happens every single time a new "free speech" platform shows up, again and again and again and again.
To maintain free speech, we will need smaller self hosted platforms that stick to a certain niche. Everything that becomes mainstream dies sooner or later.
Is anyone working on the I2P client/driver on Pascal, C , C++ , etc
Not Java Not Rust Not Python
Also how easy is it for the average person to install I2P client/driver so it can work with FTP IRC Mumble Newsgroups RSS Existing IP4/IP6 internet functionality
Installing for browsing is a matter of installing and running the daemon, and configuring any browser of choice to go over the I2P proxy. As for hosting, it's a matter of installing and running the daemon, and a bit of configuration. And optionally register a .i2p domain for free, though each user apparently has to manually add it to their address books, which is why I didn't do that.
But don't tell it to the average person, I want them to stay away from I2P. The average person is what killed the clearnet, and is about to kill Tor as well.
@fluffy@PhenomX6@BrodieOnLinux@charliebrownau@yes At least turbo-dino shit works, all the modern soydevvery always breaks, they're always either designed to constantly break, or designed by someone who has no idea how to properly code, so it'll break due to a dependency of a dependency of a dependency of a dependency getting a new major update.
@fluffy@PhenomX6@BrodieOnLinux@charliebrownau@ryo@yes I am in agreement that pascal is "turbo-dino shit," but C++ is commonly used - lots of Python libraries rely on calls to C/C++ code, and C/C++ are used for embedded problems quite a lot!
@charliebrownau@PhenomX6@BrodieOnLinux@fluffy@yes I seriously can't understand all these interdependencies sometimes. If I want to remove a fridge from my house, do I have to remove the kitchen sink, all the windows, the dining table, and the tatami mats too, because the fridge apparently depends on it?
Then again, I'm a very boomer when it comes to programming, I make stuff in PHP and C for decades, no frameworks, no dependencies, just standard code and functions I made by myself.
I like Gemini , problem is it seems 99% of the users are International Commies/Authoritarians or Self hating whites
It almost has no discussion or pages talking about - Self Defense - Homesteading - Init Fasting/Keto - Flat Earth/AE Map/Old World - Growing food/meat - Subvertd less FOSS/Linux/BSD - DIY Building Drones
I've downloaded an Arduino "keyboard adapter" that just used a built in library and dropped scan codes like crazy...and likely didn't even get enough power from the host computers bus.
@charliebrownau@fluffy@PhenomX6@BrodieOnLinux@yes The one thing that holds me back from YggDrasil is it being IPv6-based. IPv6 is already worse than IPv4 when it comes to privacy, which is why I always disable it on routers and such.
Decentralized routing inspired by the original cjdns algo. Good stuff. No central authority that can censor you. Golang, whatever, but they wanted to be productive, I don't fault them for it. They're good guys.
@cjd@BrodieOnLinux@charliebrownau@ryo@fluffy@yes I've heard people say go is better than c. It's in GCC at least and the static binary idea is cool and I've seen it used for pentesters for this reason.
I write a fair amount of go, it's ok. It's safer than C and doesn't have stupid JAWA cruft (I understand .NET is similar but never used it so can't directly say). It's not perfect, biggest failing ATM is lack of const variables.
His philosophy was about "harmful software" which wasn't just bloated webshit but also poorly designed, such as the c++ programming language, mixed in with humor you'd get it you used it.
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@fluffy@yes I know, I do recognize that he's completely right about technology, privacy, gender, and pay toilets (I don't care about the latter 2 too much, because the whole gender thing is still a tiny niche over here that barely anyone still takes seriously, and pay toilets simply aren't a thing here in Japan).
But when it comes to politics, economics, health, and so on, he's highly soycialist/pro-establishment.
That's why I said he's right about technology and piracy, but a total idiot about everything else.
I think RMS deserves more credit because he warned everyone about some bad shit and nobody took him seriously, and then the stuff he warned about started to happen, and the GNU software turned out to be incredibly useful to prevent a Bill Gates Controls Your Computer 1984 dystopia.
This guy is more just a purism extremist, and some of the stuff he's saying is just wrong.
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@fluffy@yes The thing I don't get is that there are actually people out there who legit believe that using a GPL loysense is magically going to prevent big corpo's with trillions in their bank accounts and the ability to buy courts into their favor from using other people's source code and turning it into proprietary code.
Like how war on drugs was supposed to stop drug dealers (but not politicians, surprise surprise) from growing weed. RIIIIIIIIIGHT...
@ryo >it's impossible to pick up Onion and I2P hostnames, and move them to some place else. what? just copy your privkeys to the new host.
@charliebrownau@PhenomX6@BrodieOnLinux@fluffy@yes >I dont see ((( TOR ))) as a valid secure solution oh boy this again. what's your reason for distrusting tor, now? all the "alternatives" you list don't even have the same goals as tor. zeronet isn't anonymous, ipfs isn't anonymous, opennic is an admirable project but it isn't as immediately useful as you think, i2p and tor routing are entirely different ballparks.
i've used tor since 2015, i2p for years as well, currently on yggdrasil, played with opennic, played with ipfs, and on top of that i have contributed code, hosted relays, and hosted fairly-popular hidden services on tor, so i know a lot about this area of technology and i don't appreciate when people fearmonger over how tor is insecure (it isn't) or spitball a bunch of technologies while having very little idea how they actually work or what goals they achieve
Sorry, what? Backdoors can be handy if you make specialized software for specific customers that tend to be so tech illiterate that they'll call you up all the time. Other than that, implanting backdoors is like "OH HEY CYCRIM OR GLOWIE, COULD YOU BREAK INTO MY SYSTEM ALREADY??".
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@ryo@fluffy@yes it's trivial to have logical backdoors in code, this is the wrong way to frame code safety. "safe" languages eliminate some classes of bugs, they are not safe against either a moronic or a malicious programmer
The problem with C and C++ is they're positively loaded with subtleties and gotchyas where glowys can hide backdoors. Rust and Go are much safer in this regard.
I write C, and I hate it, I'm constantly paranoid that I will mess something up. And taking a contribution from someone else is even worse, especially if they're anonymous...
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@ryo@fluffy@yes i don't see how rust or go prevent logic flaws/backdoors. the main problem with c++ is the huge amount of object oriented programing. C has a few problems, but memory safety is none of them. It's undefined behaviour, that can make and has made perfectly safe looking code vulnerable after compilation.
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@ryo@fluffy@yes are a static analyser and valgrind not part of your c toolchain? and is RCE literally the only vuln you're addressing with this thought experiment? because resource abuse, DoS, falsely returned data, and covert channels are just as serious and, in the business world, it costs just as much money to recover from downtime and data exploitation
In Rust or go, you can of course hide vulns in a contribution, but typically you would have to be submitting a patch to a security sensitive part of the code, or at least the patch would look a bit strange (referencing modules it doesn't need, extra dependencies, et .)
In C, you can typically hide memory corruption vulns anywhere that touches external data, and these are not obvious at all. Not to mention one can make their own bugs without the help of any attacker - and the attacker will just do code review and keep track of them in a private database...
@opal@PhenomX6@BrodieOnLinux@charliebrownau@cjd@fluffy@yes Oh, no idea how "it's important to have backdoors" would translate to "Rust can't prevent you from implementing backdoors", or maybe it's just my English language skills that suck (not a native speaker).
@cjd@PhenomX6@BrodieOnLinux@charliebrownau@ryo@fluffy@yes i'm aware that memory management bugs are common, i'm not denying the fact that typesafe languages fix this, but i am telling you it isn't a magical cure for bugfree code, and i am telling you that i, personally, can write and vet C code as easily as i could in a typesafe language, because i know C intimately and these classes of bugs are still easy to spot due to bad code design patterns
@BrodieOnLinux@PhenomX6@charliebrownau@cjd@fluffy@ryo@yes what isn't going to be easy to spot, *regardless of type safety*, are code logic bugs which make the code build, appear to do the right thing, but add some effectual nuance that introduces a bug and can go unnoticed way longer than memory vulns
"trivial to have" here meant "easy to include", "logical" here meaning programming logic rather than undefined behaviour within a language (accessing memory out of bounds for example)
>"Rust can't prevent you from implementing backdoors" yeah basically i meant that lol