I think it would be really cool, don’t you think the same?

Like, using Iroh + INS for a fully descentralized, uncensorable, undeletable, non-redundant, fast, secure, private, efficient, directionable CAS, then binding that to ActivityPub —or even better, ActivityPods—, so you can register to the instances that are archiving what is more interesting to you and subscribe to the archivists and topics you prefer.

Wouldn’t be cool? Or do you think it’s not as good as it seems to me?

Edit: I mean, archiving the Web, like Wayback Machine (with a similar UX) but descentralized and 100% FLOSS.

  • Hishiryo@scribe.disroot.orgOP
    11·
    3 days ago

    I don’t think so if it’s not in a single central server. Although I can be wrong, that’s not the point I mean by uncensorable and undeletable, what I mean is something like a government, agency or person can’t arbitrarily try to shut down, censor or delete something legitimate that’s archived, like has actually happened with Wayback Machine (and what’s thus one of the main reason why I propose this); of course I don’t want CSAM, doxxing, GPDR violations or illegal content whatsoever, and if the authors wants their content deleted, they obviously have the right to be forgotten, but there’s also situations when a web is totally taken down in a illegal or arbitrary way, or when something turns into lost media or almost lost media, and that’s when this is valuable, along with making the collective memory more reliable and durable.

    • Holla@feddit.org
      6·
      3 days ago

      I think with the current architecture of the fediverse it’s already quite hard for government actors and such to have arbitrary content be taken down, as any content gets federated to all the followers’ instances, where it will be stored afaik pretty much indefinitely. An exception to this would be media and external links, so an archive instance could focus on storing those as well.

      • Hishiryo@scribe.disroot.orgOP
        1·
        3 days ago

        Yes, the point with Iroh and IPFS is that they are content-adressed, and therefore addressable, so they massively deduplicate what is uploaded, and as a consequence there is hardly any redundant data, if not no redundant data, and that makes it more efficient, reducing the bandwidth and space required on all nodes, so if something is archived once, then all subsequent archives of that same thing will be just references to the original file, and will be very resilient since It’s like a torrent (you can seed, pin, or hydrate the content so it won’t disappear so easily).