1
0
Fork 0
mirror of https://code.sup39.dev/repos/Wqawg synced 2024-12-27 16:14:51 +09:00

FEDERATION.md: Write list-direct-replies proposal

This commit is contained in:
fr33domlover 2019-06-02 09:57:07 +00:00
parent 0ff5f71626
commit c7a563bd15

View file

@ -582,6 +582,34 @@ This proposal suggests a property named `events`, which maps to an
objects that aren't stand-alone and aren't actors can provide a stream of objects that aren't stand-alone and aren't actors can provide a stream of
updates. updates.
#### (8) List only direct related objects, not a flattened tree
There are various properties that typically form a tree or graph structure when
recursively traversed. And often a client may wish to fetch the entire
hierarchy. For example, there's the AS2 `replies` property. The AS2 spec says
it should list objects that are responses. But should/can that include indirect
replies, i.e. objects that are replies to replies, or should only direct
replies be listed, i.e. `replies` is the inverse property of `inReplyTo`?
In ForgeFed there's similarly a `dependsOn` property for listing a ticket's
dependent tickets, and the question arises there too: Provide a flat list
containing the whole transitive closure of dependent tickets, i.e. dependencies
of dependencies etc., or list the direct dependencies?
I suppose to some people the answer to this question is obvious, but to me it
wasn't, so I'd like to explicitly propose an answer and follow it.
`replies`, and `dependsOn`, and similar properties, map to a collection whose
items are the *direct related objects*, not indirect transitively-related
ancestors of descendants. So, `replies` is the inverse property of `inReplyTo`,
and if object A lists object B under `replies`, then the `inReplyTo` field of
object `B` should be pointing back to `A`.
It's still possible for object `B` to have its own `replies`, of course,
forming a tree/graph of discussion (and `dependsOn` forming a graph of ticket
dependencies). Whether or not those nested objects forming a tree/graph are
provided, is a separate question. See proposal B.4.
### (C) ForgeFed ### (C) ForgeFed
#### (1) Actors #### (1) Actors