• 2 Posts
  • 71 Comments
Joined 1 year ago
cake
Cake day: June 15th, 2023

help-circle
  • I’m going to give you the benefit of the doubt and assume what you said was simply confusing, but not wrong.

    So just to be clear if your raid array fails, and you’re using software raid, you can plug all of the disks into a new machine and use it there. But you can’t just take a single disk out of a raid 5 array, for example, and plug it in and use it as a normal USB hard drive that just had some of the files on it, or something. Even if you built the array using soft-raid.


  • I don’t want to sound like I’m just correcting you for the sake of it, but it’s actually important. Mastodon is the most popular right now, but Mastodon actually wasn’t around at the beginning! Before that was StatusNet, and before that was identi.ca and laconi.ca

    So those services already existed, they were the ones built for federation, and so Mastodon was started as another compatible implementation of an existing network protocol. All of that is to say that Mastodon didn’t need to make the right choices at the beginning, and they have already benefitted from this kind of network dynamic! The system has already worked once!



  • I get what you’re saying, but I think the issue with optional memory safety features is that it’s hard to be sure you’re using it in all the places and hard to maintain that when someone can add a new allocation in the future, etc. It’s certainly doable, and maybe some static analysis tools out there can prove it’s all okay.

    Whereas with Rust, it’s built from the ground up to prove exactly that, plus other things like no memory being shared between threads by accident etc. Rust makes it difficult and obvious to do the wrong thing, rather than that being the default.


  • It’s also worth saying that with cars vs transit the incentives flip. With cars I don’t want to make things too far away, but like in my previous comment I have to space things out some, and cars are good at going distances, so it’s not too bad for business to sprawl. Also, because I need parking for everyone, density of my building (like a multi-story building) requires an even greater density of parking. But parking garages are expensive, so it’s easier for me to build a bunch of single story buildings with big surface parking lots.

    On the other hand, with transit and pedestrians distances are much more significant. The goal now is to try and get as many things as possible to where the people already are. In this mode, building up is much more sensible because the more housing or offices or businesses you can put on this plot, the less people will have to walk to get there and the closer they are to prominent transit stops, etc. And if I don’t need parking, then I’m incentivized to put another building right next to this one to try and hit those same people without them taking more than 20 steps.


  • There’s a sense in which cars by their nature produce sprawl. Cars are larger than people, so if I want a building to contain N number of people, I need an empty space nearby that contains 3N of just emptiness waiting for a car, so our buildings can’t be too close together. There needs to be that buffer space between them for their cars.

    Then people have to get to the parking lots, so we need roads. But if I want N people to be able to get here, I need more than that space between our parking lots for enough cars to be able to reach me. Not to mention left turning lanes and big intersections, and of course long stretches to get past the long parking lots.

    So if we don’t have space for that in a city, we either have to knock down a bunch of buildings to make room for these things, it we have to expand outwards into the larger empty space outside the city. Which naturally leads to sprawl.

    It’s amazing to actually do a satellite view of an area, take a screenshot, and then colour in the parts that are actually a building or shop or home, and then colour all the other parts that are road, driveway, parking lot, intersection. It’s this foam which sprawls.



  • I’m not disagreeing with you, but am genuinely curious how “fairness” was counted. I feel we have a thing right now where one side will present a well reasoned, data driven, argument. And the other side will hastily throw together something based on vibes that mostly doesn’t address the issue at all. But out of a sense of fairness our current media feels like it has to present both as though they’re two equally effective tradeoffs when actually one is empty noise.

    So I’d be very curious if this system has a way of preserving true fairness without devolving into false equality in some way. Obviously nothing is perfect, but I’m curious.





  • On most modern distros (like Mint) you can do basically as much with Linux GUIs as you can do in Windows or Mac. So normal users don’t need the terminal. But if you want to do more, if you want the secret sauce, the terminal is there for you.

    But fear not! Basically all of us have some level of autism or ADHD, and the best of us tend to be the most extreme. If anything the terminal was written by autistic nerds for themselves! If you’ll be okay being a bit of a n00b for a bit, I think you’ll find there’s a lot of depth here to obsess over / hyper fixate / hyper focus on.

    There’s a reason people have been “fighting” for, like, 40 years over which terminal text editor is the superior one… The flames of war can run pretty deep, and there’s a lot of opinions.









  • I’m not the person you’re replying to, and I don’t have any videos, but I do love dumping explanation on people! So here’s some terms:

    File System: This is the way data is laid out in terms of actual bytes on the drive. It’s in charge of things like where to look to find the name of this file, or how to “last modified” date is stored, or how do I find out which files are in this folder. NTFS is a filesystem, whereas ext4 is probably the file system your linux machine is using. FAT is the older Windows one that’s still used on, like, SD Cards and stuff. That having been said File System is sometimes also used to refer to the particular data on a particular partition of a disk, like “the filesystem will allow” which really means the data on your NTFS partition. Filesystem is often abbreviated “fs”, and is in fact the “FS” of “NTFS”

    Mounting: In unix systems, such as Linux, file systems are “mounted” to a place in the folder hierarchy. Everything in unix lives somewhere under the “root” folder /, so mounting is basically saying “Okay, you want to see the files in this filesystem. Where should I put them?”, and if you say /home/user/stuff then the file “one.txt” at the root of your filesystem will now be visible at /home/user/stuff/one.txt", and if you mounted it at /mnt/things it would be /mnt/things/one.txt. The term mount is used like “attach” to mean “where do you want me to hang this new directory hierarchy on your existing one”.

    fstab: There are a few ways to mount things in modern linux. The classic is the mount command which looks something like mount /dev/sda1 /home/user/stuff which would take the device with the name /dev/sda1 and mounts it to the given path. Devices in linux usually live in /dev, and in this case are often given names like sda1 to represent the first hard drive (a), and the first partition of that drive (1). But, there are other ways! You can also click on the partition in your file browser and it will mount the disk for you, often auto-creating a mount path and cleaning it up when you’re done, so you don’t even have to think about it. Another way is fstab, which is a kind of config file that controls mounting devices. In here you can give default options for how you want drives to be mounted, and can even specify that you’d like some devices to be automatically mounted by the system on startup. This is actually an important part of how unix systems start, and how the root filesystem and other important ones get going. If you wanted your NTFS drive to always be available at a permanent location, you would edit this file to set that up. If this is something you wanted only periodically, then just clicking may be fine.

    Permissions: Virtually all unix filesystems store the permissions of files and directories as a “user” and “group” that owns the files, and then a set of whether or not the owner can “read” “write” and “execute” the file, whether other members of the group can, and then whether everyone else can. If two people were on the same computer, these would allow a person to be able to see their own documents, but not see the documents by other users. Or maybe they can see them but can’t make changes. And it also prevents random users of a system from changing important system configuration, when those config files are owned by the administrative user (called root by convention). Some config files will be read-only to normal users, and some contain secrets and so are permissioned so normal users can’t even see them. But! NFTS doesn’t follow these same conventions, so when mounting an NTFS drive on unix the driver has to produce a set of permissions that are unix-compatible, but it doesn’t have anything to work off on the disk. So the person above was saying by default it assumes the safest option is to make all files owned by the user root, and so if the permissions are the only the owner can write the files, and the owner is root, this will mean it’s effectively “read-only” to you. The terms uid and gid stand for “user ID” and “group ID”, which are the numbers that represent a user in the data. User names are basically a convenience that allows us to give a name to a uid, but it’s more efficient to store one number everywhere on disk for owner rather than a name.

    So putting it all together, what they’re suggesting is that you can use the /etc/fstab file, which has a very particular format, to specify default options when mounting your drive. These options include setting the uid option and gid option to your user’s uid and gid, so that when the filesystem is mounted, it will appear that all the files are owned by you, so you’ll have full permissions on them. They’ve assumed your uid and gid will be 1000 because that’s a common convention, but if you’re comfortable you can run the id command on the command line to output your actual uid and gid (you can ignore all the other groups your user is in for now)

    They also mentioned that when mounting you can specify if you want to mount the filesystem as “read-only” or “read-write”, etc. If you mount the whole filesystem read-only, then the write permissions stored on the individual files are ignored, basically. So if you were mounting with a command, or through fstab, you should make sure the rw option is present to clarify that you’re looking for “read write” permissions on your mount.

    That having been said, it’s possible none of that is relevant to you if you’re mounting the fs by just clicking in your file browser. One way to tell is if you right-click on some file you aren’t allowed to edit and look at the properties there should be a Permissions tab thing. And it will list the owner of the file and what access you have. If those permissions are already set to be owned by you, then this uid thing is already taken care of for you by the file browser. In that case it might be something more fundamental to the NTFS filesystem, like the locks other people are talking about.

    So those are some words and their meanings! Probably more than you wanted to know, but that’s okay. I liked typing it