Re: Let's turn aspiration into fact?

One of the reasons I like server-to-server content transfer, especially for
content that can be moderated, is that it should be possible for the
sending server to make assertions about items published, and for the
receiving server to trust those assertions to the extent that it trusts the
sending server.

This means that whether or not "original publication date", likes and
reposts, or even read counts or moderation flags, are re-used by the
receiving server, they can be used as information.

For example, if I'm moderating posts on my server and I see a request to
import 200 posts, it's hard to read them all.  But it helps if I can see
which ones were posted first and last in their original state, which ones
have lots of likes or very few likes, and use that information to sort,
prioritize and review.  After sampling the top five and bottom five posts
and seeing what the timeline looks like in overview, I can likely make a
pretty safe decision to allow this content in.  I may also make the
decision to allow this content to retain its original dates, likes,
comments (depending on licensing) etc, perhaps with an attribute that says
that this content was previously hosted at location X prior to date Y.

When the destination can trust these assertions, it really helps folks who
don't view microblogging purely as transient.  A knitter, cosplayer,
digital artist, fanfic writer or D&D homebrewer, using ActivityPub streams
to share or showcase their work, are more likely to  want their past work
to be available. (  I know folks who lost many photos completely when MySpace
lost data <https://mashable.com/article/myspace-data-loss>.  I wasn't
bothered, but that doesn't mean everybody wasn't bothered. ). Not only can
they move their work somewhere else, it appears chronologically and with
social proofs to help folks find the popular stuff.

Part of what makes some corners of the Internet rather resilient and
high-functioning is when communities of creative folks can associate their
identity with their output.  It may be pseudonymous, but it's a strong
identity nonetheless.  If you create an account on Ravelry today and look
up the account 'milele', the history there would give you a great deal of
trust that the person knows what the hell they're talking about (as long as
it's about knitting) and whether their taste runs towards normal or weird,
and whether you might safely engage in conversation with them, without
knowing at all who they are. (haha such a mystery (hint: their taste runs
towards weird and mathy)).

The weight of history of posting, words and images and projects and
contributions to conversations, is a more solid identity than any solely
cryptographic assurances.  It means that people can trust each other more,
and that people have incentive to behave in trustworthy behavior or their
slowly-building reputation is burned.  I gave a whole presentation years
ago on what caused it and what happens when this phenomenon emerged in
LiveJournal, SecondLife, DeviantArt, Ravelry, and for a while, Flickr.  But
it's also a cautionary tale because it means that if these platforms
regress or die off, those reputations built stitch by stitch or pixel by
pixel or sentence by sentence, are wasted.  I would definitely be bothered
if Ravelry lost all my projects (oh ho, mystery unveiled!)

Trusting the content history doesn't have to be perfect.  Let's say
somebody goes to the trouble of crafting a whole fake history and getting a
server to transmit that.  If their behavior lives up to their glossy
history, great.  If it doesn't, they've burned their fake reputation and
that was still a lot of effort.  Folks are pretty good at finding out when
somebody has crafted a whole fake history based on somebody else's work,
especially when their suspicion is raised.

So, to uplevel this into a vision, I have a vision where folks who view
their streams as long-lasting and treat them as archive and reputation
rolled into one,  can move them easily to a new host if the new host is
willing; and where moderators and admins , and other users consuming or
interacting, can use the weight of an account's history and even the
reactions to it, to make more contextual decisions.  The exact identity
matters so much less than the unique history of content and contributions.
That history need not be lost when services inevitably rise and fall.
Civil discourse is supported.

TL:DR; rich content history, even in a new location, even without complex
cryptographic assurances, is good for the Internet.

I think you asked for a diatribe, anyway that's clearly the way I
interpreted it!  You're WELcome!

Lisa

On Fri, Feb 9, 2024 at 12:55 PM Johannes Ernst <johannes.ernst@gmail.com>
wrote:

> On Feb 9, 2024, at 12:25, Ryan Barrett <public@ryanb.org> wrote:
>
> On Fri, Feb 9, 2024 at 12:02 PM Johannes Ernst <johannes.ernst@gmail.com>
> wrote:
>
>>
>> I don’t think this would actually be all that hard.
>>
>> But regardless, what is stopping us from doing this kind of thing?
>>
>
> I do think it would be hard. The current two core types of identity in the
> fediverse, URL ids and webfinger addresses, both have server hostnames
> baked into them. They're at the core of all of the fediverse's current
> code, stored data, and big chunks of UX and users' mental models. We'd have
> to change all of that.
>
> We have a handful of portable identity designs - Nomad, FEP-ef61, etc -
> which is great! And we may want another that's DNS-based instead of
> keypair-based; I know at least some people here would favor that. I myself
> might too, given the recoverability concerns around pure keypair-based
> identity.
>
> Regardless, even once we've settled on a specific design, migrating the
> entire fediverse to it, including servers and clients and existing data and
> users, seems like a huge undertaking. That's not stopping us, of course,
> but it's worth being clear-eyed about.
>
>
> If we went down that refactor-everything route, I agree it would be a big
> effort. I think though there are some things we can do that would be
> simpler, providing most of the benefits, specifically by building on ideas
> such as Mastodon’s https://mastodon.social/@j12@social.coop to represent
> a remote identity locally.
>
> But I don’t want to get into those weeds before we have more of a shared
> vision.
>
> I’m glad Sean laid out this set of ideas. What else?
>
> I think I’m going to start a list somewhere collecting them.
>
> Cheers,
>
>
>
>
> Johannes.
>
>
> Johannes Ernst
>
> Fediforum <https://fediforum.org/>
> Dazzle Labs <https://dazzlelabs.net/>
>
>
>
>