I am curious about #Mitra, so here are a few questions; hope you don’t mind.
(continues)
I am curious about #Mitra, so here are a few questions; hope you don’t mind.
(continues)
@feralthoughts No, this is not supported.
Say, a Mitra user boosts (reposts) a post in some thread, or posts a reply in that thread. Does the user then get notified of all subsequent posts in that thread?
Basically, is there some way for a Mitra user to “subscribe” to a public thread, or flag in some way that they would like to receive existing posts as well as get notified of all future posts in that thread/conversation?
(continues)
How does Mitra work in case of public conversations?
Say, a Mitra user writes a new public post (OP of a thread). Do they get notified of all replies in the thread, including even those reply-to-reply posts that do not tag them?
Can a respondent to the OP’s public post change the status of their reply post to followers-only or mentioned-users-only? If the OP is not a follower of that respondent, and is not tagged, will the reply be visible to the OP?
(continues)
How does Mitra work in case of public conversations?
Similar to how Mastodon and other micro-blogging platforms work. You're notified only when you're directly addressed (this is usually done by @mention).
It doesn't currently use Conversation Containers for public conversations, but I think the way notifications work will be the same when we switch to containers
Can a respondent to the OP’s public post change the status of their reply post to followers-only or mentioned-users-only?
They can reply with narrower visibility. I think narrowing of the scope should be allowed, but not widening. Although, as far as I know, Streams/Forte allow neither.
If the OP is not a follower of that respondent, and is not tagged, will the reply be visible to the OP?
Currently it is not visible.
Can Mitra instance admins subscribe to ActivityPub relays?
(continues)
Hashtags: only if they are ActivityPub actors, e.g. https://relay.fedi.buzz/ hashtags
Groups: yes
Relays: yes, and regular users can subscribe too. Actually, there is no special admin interface for subscribing to relays, we just use regular accounts.
Can a Mitra user follow hashtags and groups?
(continues)
@feralthoughts Yes, mirroring partially works, but nomadic accounts require a special command-line client:
https://codeberg.org/silverpill/fep-ae97-client
It's not possible to create a nomadic account from the web site.
I read that Mitra experimentally supports nomadic identity. Does this feature make it possible to simultaneously have the same account on two or more Mitra instances, with new activity in any of those getting mirrored on the clone accounts?
Basically, does nomadic identity in Mitra work in a fashion similar to that in Hubzilla?
@feralthoughts Server admin must enable the experimental API first.
All operations with a nomadic account require command-line client, regular interface can't be used. Cloning happens in background.
Can a user without admin access do this?
And once the CLI action is done, either by user or admin, can the user then use their account's web interface normally on either site, with no further command line interventions required? The cloning will just happen in the background?
@feralthoughts Yes, conversation containers can be used to enforce conversation scope.
Say user B replies followers-only to the OP. User C then replies followers-only to B’s post. In this case, B’s post will only be visible to B’s followers, while C’s post will only be visible to C’s followers, with further downstream replies inheriting the problem.
This is not a narrowing, it's a change of scope. Narrowing means user C can only reply to B's followers or to B directly.
To address this problem, it may be worth applying conversation containers also to public conversations? And applying them so that all replies inherit the privacy status of the OP?
Thank you for all the responses.
Personally, I think the Streams/Forte behaviour, of every single reply inheriting the OP’s privacy status, is the one that matches social expectations.
If the OP is starting a conversation, they should get to decide the contours for the entire thread, a respondent should not be able to change those. If some respondent wants to respond with a different privacy status, they can always start a new conversation and link to the OP.
(continues)
Also: allowing respondents to narrow the visibility will run into the well-known problem of each reply being visible to a different set of followers.
Say user B replies followers-only to the OP. User C then replies followers-only to B’s post. In this case, B’s post will only be visible to B’s followers, while C’s post will only be visible to C’s followers, with further downstream replies inheriting the problem.
(continues)
@feralthoughts This problem has an easy solution. Mentions don't have to be inserted into posts, they can simply be copied from the parent post.
Some platforms always used implicit addressing. I plan to do this too in the future (Mitra already uses implicit addressing in some cases, e.g. in group conversations).
So... some posters randomly start deleting the @ mentions of one or two or three or more recipients, leading to fragmented sub-threads and sub-sub-threads with random subsets of participants.
This is very counter-productive; it prevents inclusive, participatory conversations.
(continues)
If all participants in a thread received notifications even without @ mentions, this problem would not occur. In that case, @ mentions will only be required for adding new participants. Once added, they too should automatically receive all subsequent posts in the thread, without anyone having to @ mention them.
If some participant doesn’t want those notifications, they should have the option to unsubscribe from the thread.
One reason I have been asking about notifications without @ mentions: today, if lot of users join a conversation, then the @ mentions can take up 100-200 or more characters, and significantly reduce the characters available for the actual text (I know Mitra has ten times Mastodon’s character limit, but even then...). In addition, a poster may allot characters for hashtags and groups. That leaves even fewer characters for the actual text.
(continues)
Regarding notifications: I am not sure, but I think a Hubzilla OP starting a thread receives notifications of every single reply, including those that do not tag the OP. And I am far less certain here, but maybe the behaviour is similar in case of a Hubzilla user participating in a thread started by someone else?
But I could be wrong. Maybe we should ask some Hubzilla veteran. @jupiter_rowland , can you please tell us how this works on Hubzilla?
(continues)
I feel restricting all replies to OP's privacy level will prevent many possible pathologies.
But it introduces a new pathology: people who use platforms not supporting conversation containers may be not aware of the fact that their comment visibility setting will be ignored. They reply with a DM, but conversation owner forces it to be public.
So I think narrowing should be allowed.
I am not sure about the OP. Maybe OP should never be excluded.
Either way, in future, do you plan to apply conversation containers to public conversations?
Yes.
Hmm - ok.
So... if OP is not a follower of B, and B's follower-only reply to OP's public post does not tag OP, then OP won't receive B's reply? Or if B does tag OP, but C's reply to B's post doesn't tag OP, then OP won't receive C's reply? Seems undesirable, if that would be the case.
I feel restricting all replies to OP's privacy level will prevent many possible pathologies.
Either way, in future, do you plan to apply conversation containers to public conversations?
@feralthoughts Yes, it will be made simpler but maybe it will not look like Hubzilla.
Still have few doubts, but they don't matter.
For the average fedizen, the present design of Mitra's nomadic identity seems very difficult to use.
Are there future plans to make Mitra's nomadic identity design similar to Forte's? Or maybe simpler than that?
(I have never used Forte, but I am assuming that its nomadic identity design is similar to Hubzilla's, which I know can be activated from the regular web interface.)
If I understood correctly, conversation containers ensure that all replies in a thread are routed through the OP’s server on their way to other participants in a conversation. So once conversation containers for public conversations are implemented, it should be possible for at least a Mitra OP to get notified of every single reply in a thread, even if they are not tagged. Isn’t that correct?
Even mentions copied from the parent will count as characters, and eat up the limited character quota, no?
No, they can be be omitted from the text.
In this comment I will not make an explicit mention, but the message will still be delivered to you.
So once conversation containers for public conversations are implemented, it should be possible for at least a Mitra OP to get notified of every single reply in a thread, even if they are not tagged. Isn’t that correct?
Yes, that is correct.
By the way, I didn’t get the point you had made earlier:
Addressing in ActivityPub is similar to that of E-mail. A message contains a list of recipients and your server delivers a copy of the message to each recipient.
@-mentions, addressing and notifications are different things.
@-mentions are often used to determine recipients, but this is not a requirement. Similarly, applications often generate a notification when you are mentioned, but they may also generate a notification if you're addressed without @-mention.
By the way, I didn’t get the point you had made earlier:
> You're notified only when you're directly addressed ... I think the way notifications work will be the same when we switch to containers
(continues)
Even mentions copied from the parent will count as characters, and eat up the limited character quota, no?
(continues)
Irrespective of @ mentions, Mitra will always notify a user if that user is addressed in a post - would that be a correct statement?
Yes.
And a Mitra user's reply post will include implicit addressing (in future?) copied from parent post, so the user is free to delete the recipient list from the text of the message - correct?
Yes
But if the @ mentions are missing, the recipient application (Mastodon or whatever else) may or may not notify the recipients, depending on their own design – right?
Mention tags will still be added to ActivityPub object so Mastodon notifications will keep working.
Thank you for the explanations.
So...
Irrespective of @ mentions, Mitra will always notify a user if that user is addressed in a post - would that be a correct statement?
And Mitra's threading routine depends only on the sender and the recipient addresses in the posts in a conversation, not at all on @ mentions in those posts?
(continues)
And a Mitra user's reply post will include implicit addressing (in future?) copied from parent post, so the user is free to delete the recipient list from the text of the message - correct? But if the @ mentions are missing, the recipient application (Mastodon or whatever else) may or may not notify the recipients, depending on their own design – right?
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.