cross-posted from: https://l.lucitt.com/post/6770

I believe there are pros and cons for both. Imgur is great because you truly don’t have to think about disk space or bandwidth. Imgur is not great because they can delete your posts at any time without warning and leave holes on the interenet, especially if we’re talking 5, 10 , 20 years from now.

Should I invest in a beefy server to store all of my photo needs without storage anixety? Or should I just rely on a larger company to handle it for me? I think I’m already answering my own question by writing this post out, but I’d love to hear from the self hosting community.

  • Estinos@lemm.eeEnglish
    42·
    1 year ago

    One other thing to consider is that if you goal is to keep the image up for 5, 10, 20 years, there’s actually way more chances it happens on imgur than on your self hosted webserver : one day you’ll be bored of maintaining it, and that’ll be that. A good part of the web from the 2000’ was deleted because we were hosting our blogs on our own servers, until we didn’t. In an ideal world, we would have a p2p web, where content is distributed by users of the site and it stays up for as long as one person seeds it, like the Beaker Browser tried to do with dat, but this is far from what most people use today. For now, the power of self-hosting is more suitably directed at providing apps you will use for yourself than to publish things meant to stay out in the world. Unless of course you’re ready to commit to that, knowing from the get go it will be a challenge.

    • TheFrenchGhosty@lemmy.pussthecat.orgEnglish
      1·
      1 year ago

      The fact that this comment is implying that “centralization in major corporate platforma is better” with valid argument deeply bother me.

      Especially because I don’t really have a counter argument…

      Damn.

      • Estinos@lemm.eeEnglish
        1·
        1 year ago

        Ahah. Don’t worry, it bothers me as well. :) But well, we can’t solve our problems if we refuse to see them. Personally, I think the Beaker way was promising, but it was a project relying on a few people, so it died when they moved on. Maybe we will see something similar coming from IPFS, it’s a promising direction. The old recipe probably still work : for something to stick, it needs a standard and multiple implementations (that’s exactly why ActivityPub is doing so well, btw, in my opinion). I don’t see that coming from Dat, but I would bet IPFS is going there - at least it already has multiple implementations.