Latest activities
  • No activities yet ...
serving time in a domestic graveyard (known as Ohio)
kaniini@socially.whimsic.al
  • Statuses

    155

  • Following

    1

  • Followers

    0

a friendly #collectivist 🐰

destroyer of bloat @ #pleroma, #pkgconf, #audacious
slayer of techbros
previously #alpinelinux core a few moons ago and #debian much longer ago

she/her
Statuses
@rootwyrm they run Alpine and I have a fine time with them. they just take forever to come up :)
if this PDU swap doesn't fix it, i'm going to scream, it takes forever for these fucking POWER9 machines to reboot
i was, uhh, going to propose that we use the redundant power supplies on everything to you know, nicely cut over to the new PDU, but whatever
me: "hey, when can you replace that PDU?"

*everything goes down, including the PDU management itself*

me: "o... oh... okay..."
@feld @nicknicknicknick hi yes, the answer is obvious, and it even runs on the Pleroma core platform: Mobilizon
@feld @nicknicknicknick hi yes, the answer is obvious, and it even runs on the Pleroma core platform: Mobilizon
@karen @dashie @rin

(stable will always have a diverged history from develop, so when a new snapshot of develop is taken, there will be conflicts if you try to merge it. it's just how git works.)
@karen @dashie @rin

(stable will always have a diverged history from develop, so when a new snapshot of develop is taken, there will be conflicts if you try to merge it. it's just how git works.)
@karen @dashie @rin

don't merge. if you have any local changes rebase them on top of stable after switching. this will avoid the conflicts, in most cases.
i'm half tempted to just set up an icecast server that is just streaming random bullshit 24x7
if there is a higher power, i hope that they do everything in their power to prevent these graduates from touching any life-critical systems
also there are ads on the radio here for "MyComputerCareer.edu" which say they can "turn people who only know how to turn on a computer" into "IT professionals" in "4 months" and this entire proposal scares the shit out of me
ultimately, federated networks have to involve some level of trust in order to function. that's the uncomfortable truth of it.

perhaps we can define a triangle here, one between trusted peering, usability and trustless peering.
this isn't to say that OCAP or Stamps or any of the other proposals are not meaningful improvements to the protocol. they obviously are. but, increasingly, i think the architecture will never be fool-proof. as I've said before, OCAP is a mechanism to delegate trust assertions, but it does not *remove* those trust assertions.

and to build something completely safe, it must be trustless to begin with.
cryptography can be a liability when tied to stable identities (which is why I find the current security model to be questionable), so the question ultimately is, how do we build a network of pseudonymous interactions that is both meaningful on a personal level and usable?
@lanodan (although the fediverse could benefit from the underlying concept in the future, because these webfinger names could also point to datashards)
whether or not OCAP or any of these other proposals can solve the security problem depends on how we handle peers who simply choose to not care. part of the solution here is to impose new contracts, by breaking current policy.

ultimately, this is what MultiBox is about, for example -- it accomplishes the same thing as SharedInbox, but legacy implementations don't know how to use it. however -- whether or not colluding peers can abuse it by simply breaking the rules is an open question (I haven't investigated it yet): for example, if an instance maps the same inbox (or set of inboxes) to every actor in a MultiBox setup, then this is technically "valid", but...

GNU Social, for example has ActivityPub support, but it does not handle addressed collections (Mastodon's followers-only posts) correctly, thus bringing us back full circle to the original problem that lead to Mastodon adopting AP to begin with. ironic, isn't it?
if peers collude to break rules, were they ever rules to begin with? and thus, is the problem with the current state of ActivityPub security.
everyone knows @toast is really just a 200 year old matroska doll set sitting on my mantle