• 0 Posts
  • 66 Comments
Joined 6 months ago
cake
Cake day: August 15th, 2024

help-circle


  • I think you can have it, but you’d need to spend a pretty penny.

    All it would take is calling an electrician to run the appropriate wiring from the place you want the kettle plugged in to you breaker box, connect it to the breaker box with the appropriate breaker, cap off the other end with the appropriate plug (a 240V plug does exist in America), and then buy a kettle capable of receiving the rated voltage and current and splice on the appropriate plug (because I presume you won’t find one sold with that plug).

    An extremely expensive way to save maybe three minutes boiling water, but you can do it.





  • I got a 1U rack server for free from a local business that was upgrading their entire fleet. Would’ve been e-waste otherwise, so they were happy to dump it off on me. I was excited to experiment with it.

    Until I got it home and found out it was as loud as a vacuum cleaner with all those fans. Oh, god no…

    I was living with my parents at the time, and they had a basement I could stick it in where its noise pollution was minimal. I mounted it up to a LackRack.

    Since moving out to a 1 bedroom apartment, I haven’t booted it. It’s just a 70 pound coffee table now. :/



  • Technically all you need is a DNS server.

    No computer knows where <whatever.tld> is located, unless that route is hard-coded in a host file somewhere. It always has to ask a DNS server for that information. If that DNS server doesn’t know, it will probably try asking some other DNS server, and so on up a chain. Eventually, it reaches a master DNS server that either has the answer on-hand somewhere in a database, or it says, “lmao, that doesn’t exist”. All the DNS servers and your PC down the chain take that answer. They might memorize it for a little while and hand it out to anyone who asks them, but after a while they’ll ask their way up the chain again to see if the answer has changed since the last time they asked.

    In order to “create” a TLD, all you have to do is make a DNS server that doesn’t ask up the chain. Just pre-program the list of valid domains yourself. You can make them anything you want. You can even “steal” existing domains and make them point to anywhere you want. Nothing is stopping you. Your DNS server will confidently report its pre-programmed answers to anyone who asks.

    The catch is that any Internet-enabled device that you want to be able to use your fancy new custom domains needs to be configured to ask your DNS server in particular. People would have to manually set your DNS server as their master server to ask, or they’d have to set it to ask some other DNS server that is itself pointed through some chain up to your DNS server. This is an explicitly opt-in system, and getting a significant mass of people to do that voluntarily is practically impossible. But it’s not technically impossible.

    The only reason you don’t have to do this manually with every single device you buy is because most devices either come from the manufacturer with a hard-coded list of DNS servers they should trust by default, or a device on the local network whispers in their ear and tells them who the local DNS server is and the device just goes along with it. It’s still technically an opt-in system; devices are simply either already “pre-opted in”, or there’s a system running on your network that auto-opts-in every device that connects, and most devices are designed to accept that auto-opt-in the moment they detect it.

    Provided you manage to get the devices you want to listen to your DNS server, you may additionally want to set up a root certificate authority. The thing that makes the little padlock show up in your browser URL box to let you know the connection is secure. Kind of like the DNS server thing, this is also very simple–just run a cheeky little OpenSSL command or two and you can be a root CA in no time–but it suffers from the same “opt-in” problem. You have to manually configure any device you want to use your system to trust your certificates. Most devices just come with a list of “acceptable authorities” built-in, and those defaults are all most people are using. But nothing is stopping you from adding anything you want to that list at any time. You’re just limited to doing it on a device-by-device basis.

    At my company, we’ve set up our own custom DNS server and our own root CA. We serve internal websites at a custom TLD we made up, and we sign them with our custom certificates to keep the connections secure. But that only works because we’ve manually configured our workstations to ask our internal DNS server for DNS requests, and we’ve manually configured all the workstations to trust our root certificate authority. A random device that connects to our network that isn’t configured with either of those things will not resolve any of our custom domains, nor will it securely connect to them. It also breaks if the configured devices aren’t on the local company network, since the DNS server isn’t reachable from the public web. Which is fine for us, since those internal websites aren’t reachable on the public web either. But yeah, that’s an example of the limitations.

    If you want to create a TLD that will be auto-accepted by everyone who is already running the default chains of trust (which is probably what most people actually mean when they ask something like this), you have to seek out the big daddy at the root of that chain of trust and ask them to poof your TLD into existence for you. That would be ICANN, and they probably won’t do anything like that without a big fat check and a lot of corporate lobbying.

    tl;dr - The tech is built in such a way that nothing is stopping you from making your own toy, and anyone can play with your toy without needing to do much. But if you want your TLD to “just work” for everyone in the world without asking every single one of them to explicitly opt-in, which is probably what you actually want, then no, you basically can’t do that.


  • There’s lots of software out there that is available to use without payment, but is still license restricted in such a way that you are not permitted to redistribute, modify, use for commercial purposes, etc. To many, these rights are the far more important facet of “free” software, above what it costs.

    But since the English language has the same word for all of these concepts, we have all these yucks running around with zero-cost but right-restricted software wearing the “FOSS” badge thinking they’re part of the club. So some people add “Libre” to the acronym to explicitly disambiguate.



  • This is a question I see from time to time, and it’s a good question to ask.

    Your question as I understand it can be phrased another way as:

    The square root of -1 has no defined answer. So we put a mask on it and pretend that’s the answer. We do math with the masked number and suddenly everything is fine now. Why can’t we do the same thing to division by zero?

    The difference is that, if you try to put a funny mask on the square root of -1 and treat it like a number, nothing breaks, but if you try the same thing with a division by zero, all sorts of things break.

    If you define i = √-1, that is the only thing i can ever be. That specific quantity. You can factor it out of stuff, raise it to that exponent, whatever. And if it is ever convenient to do so, you can always unmask it back into that thing, e.g. i^2 = (√-1)^2 = -1. All the while, all the already existing rules of math stay true.

    A division by zero isn’t like this, because if you tried it, every number divided by zero would equal to the same thing. If we give it a name, say, 1 / 0 = z, then it would also be true that 2 / 0 = z. We could then solve both sides for zero:

    1 / z = 0

    2 / z = 0

    then set them equal:

    1 / z = 2 / z

    then multiply both sides by z:

    1 = 2

    which is a contradiction.

    i doesn’t have this problem.






  • it’s a venture capital-backed startup that has been very eager to exit its growth phase and enter its aggressive monetization phase so it can start making its shareholders some money. They’ve already tried a few things that didn’t work, like trying to turn it into a Steam competitor.

    The service to date is mostly fine. If you’re like most people who don’t mind exchanging some privacy and control for access to an app that has a nonzero professional UX design budget, it’s pretty fantastic. But the writing has been on the wall for a long time that enshittification is near on the horizon. It’s not a question of if, but how soon.