A lingua franca for content organization

AaronNGray

@[email protected] @julian @[email protected] imu

Mastodon is really an extended implementatation of the W3C ActivityStreams 2 Specification.

Regarding as:Notes and Replies as I hinted previously on my fosstodon.org account:-

The difference between as:Notes and Replies, is they are both as:Objects and they both have an as:Object.inReplyTo field that might be either a copy of the whole instance of the original Note or a URI to the as:Note. You can think of this as a normalization

This scheme is then quite versatile AFAICS. We really need to map out as many of the different protocols semantic models. Forming and providing a path for this to be performed and getting the interests and confidence of different fediverse implementations will be key.

We can and look at how we can either align them isomorphically or work out how we can produce transformations to allow conversions between different implementations.

Regarding other Fediverse systems I have not really looked at this properly as of yet, we are already getting bridging and interop implmentations we can learn from. I think relay based systems like Lemmy and others, can be treated just like feeds simular to RSS would be treated for recieving posts by subscription and for sending can be based on posting to either to named relays or to the relay of the recipient. AFAICT as the actual message structure is so simple for Lemmy anyway, there is nothing complex about this. If we can look at systems and construct archetypes then we can form a general interop model. Something like ShuttleButt might be fun but I realize there is also a lot of politics and real world issues around messaging across different networks and their communities that also have to be taken into account.

I have been looking at supporting communication between a Posts/Comments/[Nested Comments]/Likes type feed like NNTP or FaceBook and ActivityStreams. Technically AFAICS its a simple mapping very much like what NodeBB is doing nicely here, but what the users sees in the Fediverse using say Mastodon is a more disparate set of Notes (Posts / Toots) and their Replies (posts / Toots). Toot! which I use on my iPhone is nice as it forms threads of Notes and Replies. Clients are important !

infinite love ⴳ

@julian @rimu @AaronNGray a Note could be a chat message, a collection of those is a room, a higher-order collection is a guild

infinite love ⴳ

@julian @rimu @AaronNGray naturally, the nomenclature i prefer is object/context/higher-order-context

infinite love ⴳ

ah @devnull seems this self-reply never made it across to community.nodebb.org

🍀🩹一番下手・冬花 🌊🧙‍♀️

@trwnh @julian @rimu @AaronNGray embrace counting

a 0-conversation is a post
a 1-conversation is a thread
a 2-conversation is the local tl
a 3-convers

Julian Lam

@trwnh hmm.. really? I can see both, which one were you referring to?

infinite love ⴳ

@devnull ah, there might be some lag on ingesting it, then

Angus McLeod

Just a note, if anyone is wondering, that @julian and I have figured out why this topic didn't quite make it to [email protected] even though that actor is following this category. A fix will be coming soon.

Aaron Gray

@schappi @julian @rimu @trwnh can you formalize this in a document it’s a bit woolly as is

Aaron Gray

@trwnh @julian @rimu This is not exactly user friendly : /

infinite love ⴳ

@AaronNGray @julian @rimu the user friendly wording would depend on the specific presentation. there's actually no data difference between a chat message, forum post, social media status, blog entry, and so on. you could use any of a wide variety of terms at a user level.

Aaron Gray

@trwnh @julian Yes I realize we are looking at an abstract model or set of abstract models that describe real world user entities to varying degrees. But by user I meant the user of our model or models, sorry I should have been more specific there.

julian

I think that this actually makes the most sense from a technical point of view since ActivityPub is that common language which we're all speaking already.

Ironically you mentioned context and linked out to FEP-7888 which only proves my point that even within the confines of ActivityPub there isn't consensus on what a higher ordered collection would be called!

@[email protected] @[email protected] @[email protected] @[email protected]

infinite love ⴳ

@julian @silverpill @rimu @AaronNGray a context-of-contexts? nth-order context? you're right that there isn't really a term for "collection of collections", but that's really because the collection model inherits from ldp containers, which are basically like directories or folders. you can of course have subfolders or subdirectories. does this deserve a special term? well... maybe, maybe not.

infinite love ⴳ

@julian @silverpill @rimu @AaronNGray there is a tangential point here -- there is no way to signal what a collection *contains*. for example, "this is a collection of Notes" or "this is a collection of Activities" or "this is a collection of Collections" or "this is a collection of collections of collections", and so on. maybe this doesn't matter, but it could be helpful information to orient oneself...

infinite love ⴳ

@devnull the behavior i'm observing is that when i make a self-reply, it doesn't get boosted by the "threadiverse working group" actor, and for some time, it doesn't appear in-thread on nodebb. but it might get fetched later when someone else replies to it...

silverpill

@trwnh

@mikedev proposed the new collectionOf property in his recent post:

https://fediversity.site/item/01ed3c36-13b8-4a74-b111-24027bcc0697

@rimu @AaronNGray @julian

Larry Garfield

@julian Yes, please, common names for things, please. Some kind of IANA registry for object types?

Aaron Gray

@silverpill @trwnh @rimu @mikedev @julian

Interesting, will have to digest overnight against the specs.

I am fond of proper typing, but it can also get in the way if the type system is not flexible enough to express what is required.
A type system may allow disjoint unions or variants or more complex definitions required for expression.

A type annotation can be used as a hint, erased, or used as an assertion.

The FEP process can allow formalization as a route to testing in implementations.

Julian Lam

@trwnh Can you point me to the note that wasn't immediately picked up?

I can't say for certain, my guess is if it was just a self-reply (not mentioning anybody but yourself), Mastodon won't deliver it to me, since I wasn't `inReplyTo` or mentioned/tagged.