#social 2017-11-08

2017-11-08 UTC
timbl, tantek, rowan, timbl_, cdchapman, xmpp-social, _xmpp-social, KjetilK_, KjetilK and rowan_ joined the channel
#
nightpool
cwebber2: when the spec says "To do this, the server MUST target and deliver to the values of to, bto, cc, bcc, and/or audience if and only if all of the following are true", is that a double implication?
#
nightpool
That is, if the following are true is the server REQUIRED to deliver in that way? that seems to contradict the paragraph after the list, which says the server MAY filter for other reasons.
#
nightpool
If not, what is the point of saying "if and only if
#
nightpool
" there, rather then "only if" which is just a straightforward single implication?
#
cwebber2
the MAY is an escape hatch
#
cwebber2
which IIUC is pretty common
#
cwebber2
nightpool: I think the requirements definitely include the MAY normatively as an escape hatch, but
#
cwebber2
maybe it could be phrased in a better way
#
cwebber2
do you have suggestions?
cdchapman joined the channel
#
nightpool
change "if and only if" to "only if"
#
nightpool
Unless I'm misunderstaanding, and all compliant activitypub servers *should* forward things.
#
nightpool
i think in general i'm having a hard time trying to understand the exact requirements of that section and how it fits into what mastodon is doing
#
nightpool
if I get an activity that's 'cc'd to me, am I required to forward it to the person it's 'to'd to?
#
nightpool
I think a careful reading of that spec requires that I do
#
nightpool
oops, tag for cwebber2 (sorry, IRC is hard for me since I don't get notifications ^.^)
#
cwebber2
hi nightpool
#
cwebber2
nightpool: well we could lower it to SHOULD
#
cwebber2
which wouldn't require changes to existing implementations
#
cwebber2
nightpool: would that help?
#
cwebber2
nightpool: but this forwarding isn't about forwarding to individuals
#
cwebber2
it's about forwarding to one of your collections
#
cwebber2
maybe it needs a header explaining it
#
cwebber2
this is about this particular case
#
cwebber2
imagine you send a post to your followers, which include me
#
cwebber2
and I reply to your post and preserve cc: [<nightpools-followers>]
#
cwebber2
*but I don't have access* to who your followers are
#
cwebber2
maybe also your family, etc
#
cwebber2
some collection I can't read
#
cwebber2
but if I reply to you and I can't read it, you get the "ghost replies" problem
#
cwebber2
where I reply to you but your followers don't see it, then you reply to me and then other followers see that reply and don't know what it's about
#
cwebber2
nightpool: I think the big problem here is that the section doesn't explain that this is what this talking about
#
cwebber2
nightpool: IIRC that Mastodon even includes this functionality?
#
cwebber2
the forwarding-to-avoid-ghost-replies thihn
#
cwebber2
IIRC Gargron was thinking about including it at least, I assume you all did
#
cwebber2
maybe we should add a non-normative explaination of the context before the normative text
#
nightpool
we have two behaviors:
#
nightpool
- if user B replies to a post by user A, and it's to/cc PUBLIC, it gets forwarded to user A's followers
#
nightpool
er, I thought there might have been another relevant behavior, but I guess there's not
#
nightpool
This kind of fits the spec, but not really.
#
nightpool
"Server: Fetching the inbox
#
nightpool
" probably shouldn't be a common server test?
cdchapman joined the channel
#
nightpool
cwebber2: how do I turn a impl report link into a MD implementation report?
tantek joined the channel
#
cwebber2
nightpool: you don't have to
#
cwebber2
nightpool: that's deprecated
#
cwebber2
nightpool: I should delete that directory
#
cwebber2
nightpool: technically, the canonical place that they end up are now in:
#
cwebber2
nightpool: oh I replied on the ticket ot this
#
cwebber2
nightpool: let me check the "fetching the inbox" one.
#
cwebber2
nightpool: so this is technically correct... though perhaps it's an absurd part of the spec. An inbox *could* be dereferenceable to see what messages were (publicly) sent to the user, as a public collection. though I imagine probably nobody really wants to implement that.
#
cwebber2
nightpool: I'm looking at the spectext, but somewhat asynchronously
#
cwebber2
is in a meeting
tantek, npdoty_ and flackr joined the channel
#
npdoty_
trackbot, start meeting
#
trackbot
is preparing a teleconference.
RRSAgent joined the channel
#
trackbot
RRSAgent, make logs public
#
RRSAgent
I have made the request, trackbot
#
trackbot
Date: 08 November 2017
#
trackbot
Meeting: Social Web Working Group Teleconference
#
npdoty_
Meeting: Decentralized Social Networking & Games
#
npdoty_
chair: cwebber2
#
npdoty_
[introductions]
rowan joined the channel
#
Loqi
[Christopher Allan Webber] ActivityPub
#
npdoty_
cwebber2: actors can send messages to the rest of the world and an inbox to receive messages from outside
#
npdoty_
... activities can include notes, likes, shares, etc.
#
npdoty_
... not too far away from the old-school idea of MUDs etc. where actors take actions (pick up book, etc.)
#
npdoty_
... have more interactive spaces, a cyberpunk future where you participate with objects
#
npdoty
scribenick: npdoty
tantek joined the channel
#
npdoty
yay octodon.social!
#
Loqi
woot
#
npdoty
cwebber2: demonstration of mastodon, two different social network users on two different servers
#
npdoty
christopher_allen: what are the possibilities for "porting" an account to another server?
#
npdoty
nolan: only informal solutions like changing their "names" to indicate moved accounts
#
npdoty
cwebber2: Social Web Incubator Community Group is covering that topic
#
npdoty
cwebber2: private messages aren't end-to-end encrypted, just communicated between servers (transport-encrypted) and some basic trust in the administration
#
npdoty
... could use DIDs as the identifier instead of just account URLs
#
npdoty
... server-to-server side effects are specified, but that relies on server access to the contents of ActivityPub messages
#
npdoty
nolan: e2e scenario doesn't seem like the most high-value use case
#
npdoty
... relative to account export/import features
#
npdoty
cwebber2: but is still raised as a privacy issue ("advisory privacy")
#
npdoty
npdoty: one way that people have tried is to use Keybase where you sign a statement that proves that you own multiple accounts
#
npdoty
... which handles both key discovery and proof of connection of two accounts
#
npdoty
is there transparency for that?
#
npdoty
drummond: decentralized identifiers also allow you to specify those claims, including moving between accounts
#
npdoty
nolan: concern about usability, given that the competition is a streamlined centralized social networking service
#
npdoty
cwebber2: what would the user interface look like?
#
npdoty
drummond: have seen lots of demos for services like Signal where the UI really is getting to the point of being comfortable
#
npdoty
cwebber2: social networking sites can make it much easier to actually build a web of trust
#
npdoty
cwebber2: historically when people talked about blockchains for everything! zooko emphasized the utility of an entrypoint to cryptographic systems that have a human-usable experience
#
npdoty
a social network can introduce people into a cryptographically secure system
#
npdoty
contrasting to the keysigning party stereotype :)
#
npdoty
topic: games
#
npdoty
christopher: cross-game chatting or other interactions
#
npdoty
[lots of experience with MUDs and MOOs, cwebber2 gives a demo]
#
npdoty
games can include both chat and interaction with a world that you walk around
#
Loqi
LambdaMOO
#
npdoty
christopher: objects can be more complicated, respond to different kinds of verbs, etc.
#
npdoty
cwebber2: these games were typically community programmable, users could build objects that would interact
#
npdoty
nolan: in a federated system, how would you prevent cheating?
#
npdoty
Lessons from Habitat, and a 1980s game Habitat, that had collaborative interactions, that used consent/capabilities
#
npdoty
Second Life has similar techniques
#
npdoty
could use VRML as well
#
npdoty
smart contracts to enable secure capabilities
rowan joined the channel
#
npdoty
sending messages from actor to actor about your position, etc.
#
tantek
is glad to see this channel being used for this :)
#
Loqi
tantek: ajordan left you a message 1 day, 19 hours ago: just read through the logs, very interesting! glad y'all had a good/productive time and I just love "hey we had a meeting!" on that page :-)
#
tantek
:) ajordan
#
npdoty
federated environments, could involve exchange of currencies between them
#
npdoty
how much of this is possible to implement with ActivityPub today?
#
npdoty
cwebber2: we could have customized software that had existing Mastodon-style feeds in some columns and have another column that is the game, with updates to the game state
#
npdoty
... the effort to build that could be a compelling use case for adoption of federated social networking
#
npdoty
[cryptocurrency conversation]
#
npdoty
next steps?
#
npdoty
cwebber2: interest in trying out game development next year
#
npdoty
webrtc and data channel for peer-to-peer connections
#
npdoty
[adjourned]
#
npdoty
rrsagent, draft minutes
#
RRSAgent
I have made the request to generate https://www.w3.org/2017/11/08-social-minutes.html npdoty
rowan joined the channel