03.05.21
===

## name jam

- hash
- keypair (w. author inside)

message
- message
- message_encoded

entry
- seq_num
- entry_signed
- entry
- log_id
- encode, decode, verify ..

schema
- persona
- ...

wasm
- ...

## materialize jam

struct Message {
  fields: MessageFields,
}

message:
  fields: {
    message: Text
    date: Text (ISO date string)
  }
  
struct Schema "chat" {
   fields: ... name, type
}

- aquadoggo:

registratoring schema

1. schema -> CDDL (create_cddl)

CDDL in memory

CDDL validates message payload -> yes / no

2. create tables

create sql for tables (create_sql)

CREATE TABLE IF NOT EXISTS cat(actually the entry hash, and not a cat but a chat) (
    message         VARCHAR(64)    NOT NULL,
    date            DATETIME       NOT NULL,
);

- incoming new entry using the chat schema
- check bamboo integrity
- check message schema via CDDL
& check "chat" schema via CDDL
- create row in "entries" table
- create task for background materializer
- tell sender it's alright go to bed, and send next entry arguments in case you can't sleep

- materializer background job handles task
- receives notification about new incoming entry (coming from replication or creation)
- add incoming entry to worker queue
- create or update row in cat table

The raw data
==================

entries
 -
     - author: "andreas"
     - log_id: 1
     - seq_num: 1
     - message: {
        hash: "ABC"
        action: create
        schema: "Chat"
        fields: 
           message: "Hello"
           date: ...
     }

logs
  -
      - author: "andreas"
      - log_id: 1
      - schema: "Chat"
      - seq_num_head: 120
      - materialized_seq_num_head: 20 

The (materialized) DB
=====================

chat
  -
      - id: "ABC"
      - message: "Hello"
      - date: "..."

RPC: panda_query "Give me the first 10 chat messages sorted by date

---

### Relations?

**First scenario (same schema)**

NODE 1
hash author   message   reply_to
1    alice    hi        -
2    bob      hi back!  1

NODE 2 (didn't receive alice message yet)
materializing this:
2    bob      hi back!  1

**Second scenario (two different schemas)**

One schema depends on another schema - node should understand that the second one should be materialized as well

schema chat messages
hash message   profile_name
1    hi        2
2    hi back!  1

schema profile names
hash  name
1     bob
2     alice

**Third scenario (writing to the same data instance by multiple authors)**

Two different key pairs ("devices"):

Persona:
- panda
- shirokuma

entries w. messages
hash author    action    fields                       id    seq_num
1    panda     CREATE    { profile: "Cat" }           %     1
2    panda     UPDATE    { profile: "Dog" }           1     2
3    shirokuma UPDATE    { profile: "Kuro" }          1     1
4    grizzly   UPDATE    { profile: "I HATE CATS" }   1     1
😾 > boo, grizzly! You are not allowed to do that update!

materializing this:
hash  profile  
1     Dog      

Do we need a `Permission` or `Collaboration` system schema?

---

> Automerge dealing w. conflicts: https://github.com/automerge/automerge#conflicting-changes
> SSBs approach to "Personas": https://gitlab.com/tangle-js