nip17: "seen" event.

17.md

@@ -8,7 +8,11 @@ Private Direct Messages
 
 This NIP defines an encrypted direct messaging scheme using [NIP-44](44.md) encryption and [NIP-59](59.md) seals and gift wraps.
 
-## Direct Message Kind
+## Message kinds
+
+Each of these message kinds can be gift-wrapped as described afterwards. Clients have to first unwrap them to then see their kinds.
+
+### Direct Message Kind
 
 Kind `14` is a chat message. `p` tags identify one or more receivers of the message.
 
@@ -31,7 +35,7 @@ Kind `14` is a chat message. `p` tags identify one or more receivers of the mess
 
 `.content` MUST be plain text. Fields `id` and `created_at` are required.
 
-An `e` tag denotes the direct parent message this post is replying to. 
+An `e` tag denotes the direct parent message this post is replying to.
 
 `q` tags MAY be used when citing events in the `.content` with [NIP-21](21.md).
 
@@ -41,7 +45,7 @@ An `e` tag denotes the direct parent message this post is replying to.
 
 Kind `14`s MUST never be signed. If it is signed, the message might leak to relays and become **fully public**.
 
-## File Message Kind
+### File Message Kind
 
 ```jsonc
 {
@@ -59,7 +63,6 @@ Kind `14`s MUST never be signed. If it is signed, the message might leak to rela
     ["decryption-key", "<decryption-key>"],
     ["decryption-nonce", "<decryption-nonce>"],
     ["x", "<the SHA-256 hexencoded string of the file>"],
-    // rest of tags...
   ],
   "content": "<file-url>"
 }
@@ -82,6 +85,46 @@ Kind `15` is used for sending encrypted file event messages:
 
 Just like kind `14`, kind `15`s MUST never be signed.
 
+### "Seen" event
+
+An event of kind `30016` that CAN be emitted automatically by a client whenever a chat window is displayed to the user, even if the user hasn't acted on it.
+
+```jsonc
+{
+  "id": "<usual hash>",
+  "pubkey": "<sender-pubkey>",
+  "created_at": "<current-time>",
+  "kind": 30016,
+  "tags": [
+    ["d", "<subject>"],
+    ["seen", "<latest_message_id>", "<previous_id>", "<...>"]
+  ],
+  "content": ""
+}
+```
+
+- `d` must be set to the conversation subject, or it can be empty or omitted in the most common case of a `<subject>` not existing.
+- `seen` must be set to the id of the last messages seen. It can contain any number of ids, ordered from the newest to the oldest.
+
+This event SHOULD be discarded whenever a new event is received for the same conversation.
+
+Any messages with timestamp before the last item in the `seen` array are assumed to have been seen.
+
+If there is a gap in the `seen` array that indicates a message may have been missed for whatever reason.
+
+For example:
+
+- `alice` sends message `aaaa`
+- `alice` sends message `bbbb`
+- `alice` sends message `cccc`
+- `bob` sends event with `["seen", "bbbb", "aaaa"]`
+- at this point `alice`'s client should display `aaaa` and `bbbb` as seen, `cccc` as unseen
+- `alice` sends message `dddd`
+- `alice` sends message `eeee`, which is lost due to relay malfunction
+- `alice` sends message `ffff`
+- `bob` sends event with `["seen", "ffff", "dddd", "cccc"]`
+- at this point `alice`'s client should display all messages as seen, except for `eeee` which should be displayed with an error indicator
+
 ## Chat Rooms
 
 The set of `pubkey` + `p` tags defines a chat room. If a new `p` tag is added or a current one is removed, a new room is created with a clean message history.

73.md

@@ -20,7 +20,6 @@ There are certain established global content identifiers such as [Book ISBNs](ht
 | Geohashes              | "geo:`<geohash, lowercase>`"                               | "geo"                    |
 | Movies                 | "isan:`<id, without version part>`"                        | "isan"                   |
 | Papers                 | "doi:`<id, lowercase>`"                                    | "doi"                    |
-| Music pieces           | "isrc:`<id, uppercase>`"                                   | "isrc"                    |
 | Hashtags               | "#`<topic, lowercase>`"                                    | "#"                      |
 | Podcast Feeds          | "podcast:guid:`<guid>`"                                    | "podcast:guid"           |
 | Podcast Episodes       | "podcast:item:guid:`<guid>`"                               | "podcast:item:guid"      |

F3.md

@@ -1,48 +0,0 @@
-NIP-F3
-======
-
-Scrobbling
-----------
-
-`draft` `optional`
-
-This NIP defines `kind:1073` as a _scrobble_ event. It has the following format:
-
-```yaml
-{
-  "kind": 1073,
-  "tags": [
-    ["i", "isrc:<code>"],
-    ["title", "<song title>"],
-    ["album", "<song album>", "<optional variant>"],
-    ["artist", "<musician or band>", "<optional variant>"]
-  ]
-  // other fields...
-}
-```
-
-### Metadata
-
-`"title"`, `"album"` and `"artist"` should always be included, but in case the ISRC code is also included (as it should unless it's too impossible) ISRC can be seen as the authoritative source for all metadata related to the given song, the other tags being just hints.
-
-### Relays
-
-Scrobbles should be published to the author's "write" relays (as per [NIP-65](65.md)).
-
-### Usage
-
-Given that Nostr is open it makes no sense to speak of a _global_ "most listened" tracks, for example, as _last.fm_ does.
-
-It does make sense, however, for clients to display the "most listened" and other stats in the context of individual relays or groups of relays, as long as such relays are gated to a certain community or trusted to not have sybils or spam.
-
-It also makes sense to display data such as "most listened by your friends", "most listened by friends of friends" and other arbitrary selections.
-
-The same is true for things like a feed of "recently listened" tracks.
-
-#### Commenting on tracks
-
-It makes sense to comment on a track by using a [NIP-22](22.md) with a `"i"` tag containing its ISRC code, as specified in [NIP-73](73.md). That way it's possible to display comments in context with scrobbles on a track page, for example, and whatever else.
-
-#### Correcting History
-
-It shouldn't be hard to "edit" an old scrobble by simply deleting it and creating a new scrobble.