• 1 Post
  • 33 Comments
Joined 1 year ago
cake
Cake day: July 21st, 2023

help-circle











  • BeOS went under.

    Ed: I was a huge apple fan, bought an apple clone from Power Computing. Then Apple revoked the licensing that allowed all the apple clone companies to exist. That’s when I went to BeOS which would run on my clone, and got a multicore intel machine too. When BeOS went under I tried Suse. Had kind of a sucky UI in my opinion, but I hung in there with linux as an alternative to windows and went Ubuntu/Debian/Arch/Nixos and I’m still on nixos now. Its pretty much my exclusive OS since I quit my job that required windows 5 or 6 years ago.






  • I’d leave the sorting up to the users. So for a post where 5 users tagged it as ‘baroque music’, and 5000 tagged it as ‘boring’, one could sort the feed by ‘total tags’ on a post indicating general interest (5005), or just ‘total tags I follow’ which might just be ‘baroque music’ (5). Or maybe reverse sort by tags so ‘boring’ stuff is towards the bottom.

    I’d think that ignoring tags would be a thing for users, so “libtards” or whatever could be ignored.

    Tag mods could ban problematic users, so someone could get banned for tagging ‘corporate lies’ where the mods think it doesn’t belong. Offenders could make their own ‘corporate liez’ tag though, I suppose.

    A tag hierarchy might be desirable, like everything tagged ‘baroque’ also getting ‘music’ automatically. Perhaps through agreements with the mods of each tag.

    ‘archive of our own’ I’ve heard has a solid system of tag moderation. Not sure it would be appropriate for a system like this.


  • What I think would be interesting would be a link aggregator based around tags rather than subcommunities. Moderation would be based around these tags. Your feed would be based on tag queries. Posts could have multiple tags, assigned by the original poster or by users. Assigning tags would have a similar effect to voting, so a post might get tagged by 1000 people as ‘corporate lies’, or as ‘music’, or whatever else.

    Nice thing about this would be finer grained queries with news, for instance. Could get ‘politics’, but minus ‘corporate lies’.



  • I think nixos is still niche, but seems to be gaining momentum. It has some unique features:

    • Every package has its own dependencies, so you can install a 7 year old firefox alongside the latest, and have no interference.
    • Packages with dependencies in common still share them (for space savings).
    • Abandons the HFS, but can still fake it for apps that need it.
    • Can make dev environments that are exactly reproducible across machines, and only exist within a specific shell session. So you can have a project that relies on an out of date version of a compiler, and another that uses the latest, and run both at the same time.
    • Make your own packages that other people can install using a git repo address.
    • The package language can also describe a machine’s configuration; systemd services, default packages, user accounts, etc.
    • You can build and remotely deploy a machine config in one line.
    • You can cross compile a machine config for another cpu architecture, like ARM.
    • OS upgrades are atomic, and reversible. If it doesn’t work out, you can go back to the previous config.
    • No reason to ever reinstall. Recently upgraded a machine that had sat in a closet for 5 years to the newest release. Flawless upgrade.
    • Nixos boasts more packages than any other distro, over 100,000.

    There are certainly downsides - poor docs, confusing core language. Instructions for installing something on say debian will not work on nixos. I do think this style of package management is the future, if perhaps not this specific implementation. It can be a pain but its also super solid.