Hi Fediverse denizens. I've been working on a project I hope will help Fediverse devs make software that federates across ALL services, not just Mastodon-plus-a-few-others.
-
Hi Fediverse denizens. I've been working on a project I hope will help Fediverse devs make software that federates across ALL services, not just Mastodon-plus-a-few-others.
It's called the Fediverse Schema Observatory and before launching it, I am opening up a public comment period re: safety and privacy.
This post lays out what the software does, what data it records, and why I think it's safe to deploy:
Request for Comments: The Fediverse Schema Observatory
Darius Kazemi introduces the Fediverse Schema Observatory, which enhances the interoperability of Fediverse, and seeks public feedback on the tool.
Applied Social Media Lab at Harvard University (asml.cyber.harvard.edu)
Please give feedback here or at the email in the post!
-
DJ Sundog - from the toot-labreplied to Darius Kazemi last edited by
@darius hey, this looks solid to me! what's the opt-in process look like?
-
Darius Kazemireplied to DJ Sundog - from the toot-lab last edited by
@djsundog Asking a very specific version of this question because I want to make sure what I wrote came across:
can you describe to me what you think opting in will do? (I ask it this way because I wanna QA my blog post haha)
-
@darius interesting! thanks, we'll dig in when we can
-
@darius I LOVE the idea of a schema observatory. I’ve done quite a few projects in the past that involve schema registries, and they can be so helpful to a community who cares. This idea of effectively reverse-engineering the current utility of supporting existing federated schemas is awesome.
-
Here is a video demo of the tool if you just wanna see what it's about in 4 minutes.
-
DJ Sundog - from the toot-labreplied to Darius Kazemi last edited by
haha, fair enough - based on my first (fast) read-through, I think opting in will send all public (or in your parlance, hyperpublic) posts to your server for aggregation into the dataset.
I'm guessing that the opt-in process will look a lot like subscribing to a relay that just happens to never send any traffic back to my server
-
Darius Kazemireplied to DJ Sundog - from the toot-lab last edited by
@djsundog Precisely yes and yes! You'll just add it as a relay
-
@darius Your stance on semi-anonymized data, the idea of using relay traffic exclusively, the (forthcoming) web presence of being able to browse this information: wonderful.
-
@darius you start to get at version interop at the end of the video, and separately you mentioned in another comment asking “what would opt-in look like?” Any gluing of these two together, as in a registration of schema transformations from one to another? So far it seems like thenproject is saying “ok, now you as a fedidev can target the right schemas for your community that will deliver value,” but it leaves the outliers as just that.
-
The Psychotic Network Ferretreplied to Darius Kazemi last edited by
@darius JESUS MARY AND JOSEPH!
We do NOT need easier access to RESULTS.
We need better documentation to the constructs that deliver those results.
At a certain point, you have to start admitting that layers of extrapolation are counterproductive.
The Fediverse is already a tangled mess of RESTFUL calls.
-
@darius Maybe there are already projects that do this, and discovery is enough of a catalyst? Maybe everyone just falls back to a basic `note`? Forgive my ignorance, I’m basically just rehashing things I’ve had to work through previously.
-
@everyplace I mean I am hoping that people can build stuff like that on top of this! I just was gobsmacked that the basic data wasn't even out there really.
Also the schema data has already proven useful in standards discussions
-
Darius Kazemireplied to The Psychotic Network Ferret last edited by
@nuintari I don't really understand what you mean by "easier access to results" here
-
The Psychotic Network Ferretreplied to Darius Kazemi last edited by
-
@darius what happens when something you can post here will get strikes on other platforms? Some are worse than others but I also think it's a great idea if you can get around actually free speech!
-
Jennifer Moore 😷replied to Darius Kazemi last edited by
Seems to me like a cool and potentially very useful thing!
Also, respect for the lovely job done here of explaining it up front before people have a chance to get worried - which "ought to be" common sense, but in practice seems to be remarkably uncommon sense
-
a: this is a super valuable tool, very helpful, it's great youre building it!
b: I genuinely do not think the fediverse can exist if we're taking the idea seriously that surfacing activitypub data formats requires an opt-in consent. Taking that idea to its logical conclusion means that its literally impossible to build a federating AP server.
-
@YakYak_OG @darius my understanding is that as the software is just recording the shape of the protocol level messages, the actual contents of the messages (the free speech) is completely irrelevant.
i.e., it doesn't matter what the contents is or who said it because the software ignores that. It records “<string>" or “<uri>” or whatever.
-
@darius I think this is going to be fundamental as a tool for compatibility between Fediverse software as the network grows. Understanding what is being sent, schema wise, is essential to being able to handle it.
I also wouldn't mind running this as a proxy in front of a server, where I can submit information about activities I receive, but anonymizing everything before forwarding to the collector.
