Join GitHub today
GitHub is home to over 31 million developers working together to host and review code, manage projects, and build software together.Sign up
Support ActivityPub federation in Talk #2078
With respect to commenting use cases, my fascination here started when I was first PMing at Livefyre in like... 2011. We had all these bloggers using the free version of the product on their blogs, and those same bloggers were very active commenters on their friends' and peers' blogs. And they thought a lot about comenting on others' blogs as a way to drive new visitors to their own.
So if I'm reading an awesome blog post or article that someone took a lot of time to work on, and it motivates me to also invest some time and energy in leaving a thoughtful (maybe long) response, why should I post that into the comment system below the article (where it could be censored/moderated, deleted, buried after a 'load more' button, ad tracking by the publisher I have no way of knowing about, I maybe have to agree to T&Cs, give up authorship rights, etc), instead of posting it on my own blog?
If my blog were ActivityPub-enabled, then I could explicitly post a blog post 'in reply to' another blog post (or coral project comment). My server would then do some invisible ActivityPub-handshaking with the parent post's host server (e.g. coral talk), and notify it that my reply exists at another, external URL.
After this, it's kind of up to each community how they want to moderate and display these external comments. They could treat them just like 'local' on-site comments, perhaps after showing up in a moderation queue. Or they could block them entirely.
You can see a prototype of this sort of thing with:
Also, practically, distbin is open source and compatible with the JS/node.js ecosystem that Coral Talk also is implemented in. So there may be room for piggy-backing on the same modules, tests, etc that I implemented over there.
See also this from last time someone at @coralproject asked me about this:
Think about budgeting for it in 2019 planning. I'm available for hire to implement. Or happy to spend an hour or two advising pro bono. I want Talk to be as awesome as possible and become a top app on permanent.cloud.
I posted a comment on this blog post on Mozilla Hacks about ActivityPub to this issue so others can weigh in. Maybe we'll get some volunteers!
I don't see the evidence to substantiate that it's any closer to Webmention than ActivityPub.
Webmention is like a less expressive, less used, non-JSON-based alternative to ActivityPub Delivery, with no extension process (and no extensibility-via-linked-data, which doesn't require a central registry and process around it). Other projects have chosen not to implement it at all (though I wish they would have. The more interop the better).
The W3C Social WG was chartered to build a federation protocol making use of a JSON-based social data syntax. Webmention doesn't use JSON. IMO it was was clearly out of scope of the WG charter, but no one belabored that point. It was useful to standardize in addition to ActivityPub mostly for political reasons.
But sure, Talk could still be compatible with federated replies that speak Webmention but not ActivityPub. In practice, when a piece of software gets the step of 'I need to notify resource X', it could first try to use ActivityPub. If resource X didn't accept that, the implementation could then fall back to Webmention. The core required to implement ActivityPub could be 'triggered' by Webmention too.