Well, a mastodon user will see it in the context of a timeline, and timelines are generally limited. On mastodon, by default, a total timeline doesn’t exceed 400 posts. If you are looking at the original, you are looking in Discourse anyway. So while it’s true in theory I don’t see it causing actual confusion in practice. You have to already be a follower to see the content; following a link to the original takes you into Discourse where the confusion is resolved.
Not perfect, but perhaps a “least bad” option?
I suppose, as an alternative, you could federate out an edit that annotates the post as superseded by a transfer of ownership, kind of like adding the “discuss this on our forum” link?
Sure, I’m not arguing it’s not a difference, only saying that it seems like an acceptable difference, compared to blocking a useful and utilized Discourse feature.
Deleting the original will orphan threads in the context of other platforms to which you are federating, since you can’t change the actor associated with an activity in an edit (as I understand it).
An alternative might be to stop federating edits at all if the post in Discourse has a different author from when it was first federated. Maybe with a warning? “Changing owner will disable federation for this post, do you really want to proceed?”
Yes. That is what it looks like in Discourse to a normal user, who might not know that they can click the pencil icon and go through changes to review, or doesn’t have permissions. These are ingrained in normal Discourse use, as I see it:
- Making a post a wiki is saying that you accept others’ edits appearing in normal use under your own name
- Even if it’s not a wiki, sufficiently-privileged users can edit your post in Discourse, depending on site configuration
From my perspective, this is as close to equivalent as you get, given the differences in underlying model.
As I see it, “click to see post on original site” already shows different content between implementations that are fediverse-first, even ignoring this plugin. Different set of comments visible, different markup, different handling of articles. So some differences being visible through this plugin does not surprise me; I think it’s inevitable.
(Thank you again for your thoughtful consideration of these ideas. I recognize that these are hard boundary cases, I’m grateful for the work, I do not assume that my ideas are best, and I don’t mean to create any sense of obligation for this phase or any phase of work on this, or even to respond to anything in particular that I’m writing.)