Giver of skulls

Verified icon

  • 0 Posts
  • 763 Comments
Joined 101 years ago
cake
Cake day: June 6th, 1923

help-circle

  • Someone needs to curate and maintain the blocklist. Paying for software once is a surefire way to have it stop blocking anything a few years down the line because the company stopped updating their lists when the sales ran out.

    If you pay a set fee for something that requires continuous maintainance you’re either overpaying enough to compensate for cash flow interruptions or you’re at risk of the software being discontinued. The latter is especially true if you’re getting stuff for free, but in that case you’re also not entitled to any service.

    As for “why not block teamviewer at the router”: because then anydesk will still work. And if you block anydesk, Microsoft Remote Support will work. And if you block Microsoft Remote Support, RustDesk will work. And if you try to block every single RustDesk server, grandpa’s internet probably no longer works anymore, because you’ve just blocked off every data center in the world.




  • You’re not doing AJAX without Javascript, and that’s what the Google search site is optimised for. Plus, there’s no way to deal with the mandatory cookie consent popup without additional page loads either.

    You can do most of Google with CSS but you can’t do it easily without sacrificing functionality and Google doesn’t care about the people without Javascript anyway. Why invest time and effort into making this stuff work for customers that don’t earn you anything? It’s not an open source nonprofit that cares about its users, we’re talking about Google.



  • They need Javascript to serve users an experience that doesn’t look like it’s from the 90s. “You don’t need Javascript” is technically correct in the same way you don’t need Google because you can go look through an encyclopedia in the library.

    The kinds of people that disable Javascript probably don’t use Google anyway, and if they do, they’ll have their browsers so full of tracking protection that serving them costs more money than it earns.




  • Exactly. Everyone wants the cheap and easy solution when something breaks, but nobody wants to pay the price for the cheap and easy solution to be available upfront, because what are the chances they run into a problem like that?

    In this specific case, there is a credible ulterior motive for the company not to make cheap repairs available: the government will pay the bill if they sell a new expensive product and all the training/rehabilitation that comes with it. On the other hand, there is a very valid reason why things like batteries are so expensive to replace and why you can’t find replacement batteries for a lot of products a certain amount of time after production ends.




  • Because there is little difference when it comes to passively degrading components like batteries. You can’t produce a battery and leave it in storage for a decade, the battery will degrade on its own. The only way to keep reserve batteries is to keep producing them, and maintain a production line for all that time. That’s prohibitively expensive for small markets like these.

    A relatively simple solution is to stick with batteries that have a standard shape and size, but it’s not like you can just stuff a button cell in there, you need more power to operate the controller chip.

    It’s pretty shitty that the company didn’t produce a backup controller box that works without having to stick to the wearable watch form factor that just takes a bunch of rechargeable AA batteries, but you can’t expect what is essentially a smart watch to still have accessible replacement batteries in twenty years.

    This isn’t exclusive to medical devices, either. Computers running DOS or Windows 95 are still operating millions of dollars of machinery and are slowly failing and collapsing over time. The amount of affordable replacements (even at an industrial level) is slowly starting to dwindle. Nobody is producing floppy drives anymore, nor new floppies for that matter, so if that industrial controller you bought in the early 2000s dies you have to hire a computer greybeard to fix your hardware or replace the entire system.

    In my opinion, it should be put into law that once a company stops supporting their bespoke hardware, the copyright and patents protecting them should expire immediately, so that once a company drops support anyone else can pick up where they left off.

    However, anything with a computer in it has a limited lifespan, and that lifespan is significantly shorter than that of a human being. Even with the code and blueprints publicly available, someone still needs to find the compatible hardware, alter the designs to operate on modern commodity hardware, or pay a factory to ramp up a production line if they have the million(s) to do so.


  • Arbitration court with one person is a win for the company. Arbitration court with a thousand people is a massive loss for the company. That’s why these arbitration clauses aren’t always bad. If anything, for small cases they’re good for the people because the bulk of the legal charges are paid by the companies that write these clauses.

    A bunch of large companies went through a phase where they all went for arbitration clauses, and a bunch of them moved back quickly after they found out how much more expensive paying for ten thousand arbitration cases was compared to just one single class action lawsuit. Maintaining ten thousand legally binding, individually composed outcomes can haunt them for decades if they’re unlucky.

    Steam has learned the same lesson here.





  • DOCSIS 3.0 is a 2006 spec that provides less than a tenth of the bandwidth of DOCSIS 4.0. With the way channels are redistributed, you may not even get more than 100mbps/10mbps if you plug in your DOCSIS 3.0 modem, no matter what your subscription may be, depending on how your ISP deals with old hardware.

    The cable frequency spectrum is shared with everyone else, and your ISP isn’t slowing everyone down to make your hardware work, so you’re pushed into thin channels with limited bandwidth that others can use to pull 10gbps down and 6gbps up while your modem will struggle to get any decent speeds.

    In theory your ISP could be tolerant to old modems and redistribute their channels such that you’re getting the full speed, but that does mean your entire area gets a lower combined total network speed when people do buy newer hardware. Older modems waste network bandwidth so in congested areas the other side may allocate fewer channels to them.

    The latency did improve significantly between 3.0 and 4.0 (ten years of development will do that) but it probably won’t be your biggest problem.

    As for the WiFi, I’m still on 802.11ac and I don’t really care that it can only do 520mbps down on my devices. There are some latency advantages to newer WiFi as well but they’re pretty inconsequential if you don’t replace your old modem.

    As with so many things, you can give it a go and see if it works. If your performance is not sufficient (or your ISP doesn’t even let your modem connect) you may need to invest in a newer modem.


  • Newer DOCSIS standards allow for using more frequencies for both upload and download as well as newer techniques. If some frequencies on a network are reserved for 3.1, the frequency space for 3.0 will go down and so will your performance. The frequency space that used to provide 125mbps can now provide someone else with several hundreds of megabits, so you get kicked down to 75mbps for everyone else’s benefit.

    DOCSIS 3.0 came out in 2006 and 3.1 in 2013, and 3.1 has already been succeeded by 4.0 in 2017.


  • The thing with Windows is that the three magical commands (sfc, that DISM tool, fixboot) will usually fix most weird OS problems. To the point where any Windows troubleshooting session should include either the results of the first two, or instructions to use them.

    Once SFC and DISM can’t fix your install, you reinstall Windows. There are alternatives, but if you’d know them you wouldn’t be asking random Windows users on a forum. You can figure out a lot by enabling various tracing and logging features, listing open file handles and tracking file system calls, but the moment you need to take out sysmon you’re either in for a weekend of troubleshooting or wasting your time.

    Similarly, there are oneliners for Linux that’ll reinstall every package installed on the system and that has helped me recover my broken systems several times.