From 95ff6371303ac28d05b25fd9f6e436c5d0a58d4c Mon Sep 17 00:00:00 2001
From: Yigit Sever
Date: Wed, 7 Apr 2021 01:42:44 +0300
Subject: Include big picture

---
 README.md | 24 ++++++++++++++++++++++++
 1 file changed, 24 insertions(+)

(limited to 'README.md')

diff --git a/README.md b/README.md
index 920d594..11e1861 100644
--- a/README.md
+++ b/README.md
@@ -13,6 +13,30 @@ $ curl --location --request POST 'localhost:8080/transaction' --header 'Content-
 }'
 ```
 
+# how?
+
+## authentication
+Students generate their own `keypairs` and authenticate with their METU Student IDs.
+Authenticated students propose transactions, between them and another node (=public keys) or between the grader (=bank) and themselves.
+
+## transactions
+Transactions are `signed` using the proposers private key.
+(This whole public/private key + signing process will require some crypto dependency, **todo**)
+
+## blocks
+Blocks are proposed using `N` transactions, this can be an exact number (=20) or if the last block is *some time* old then small blocks can be proposed.
+Block proposal: `Block` + some `nonce` is hashed using a *simple* hash function, resulting hash should have some property that will require some computation time (~1 minute? 10 minutes?) to find (=guessing) Proof-of-work scheme.
+First proposed valid block is accepted, if assertions hold.
+(No consensus, we are the sole authority, there's no blockchain here, only a glorified database and busywork)
+
+## payment
+First transaction in the block is called *Coinbase*, the block reward is paid to the *output* (Bitcoin notation, different) of this transaction.
+If we do this then the rest of the transactions are just make believe playing.
+So banker + block reward approach seems better.
+
+## then
+After the new block, stale transactions are cleared?
+
 # Big Thank List
 - https://github.com/blurbyte/restful-rust
 - https://github.com/zupzup/warp-postgres-example
-- 
cgit v1.2.3-70-g09d2