# Pastebin 41mFs3Qz
Hey Guys, was trying to understand how stellar works from the implementation perspective. Here is what I have understood. Can some validate this :

- Create tx in the form of XDR in stellar lab
- Call route 'tx'
- XDR gets parsed into `envelope` 
- Get the hash and Check if its already present in the txPool (i.e `mPendingTransactions`) otherwise send response `TX_STATUS_DUPLICATE`
- Check if tx is valid otherwise send response `TX_STATUS_ERROR`
- Check if balance is sufficient othewise send response `txINSUFFICIENT_BALANCE`
- Add tx into `mPendingTransactions` and send responses `TX_STATUS_PENDING`
- Form a `StellarMessage` - msg, with `type=TRANSACTION`, `transaction=envelope`
- broadcastMessage(msg)

When the next ledger starts:

- Copy from `mPendingTransactions` to `proposedSet`
- Remove invalid from `proposedSet` // proposedSet->trimInvalid(mApp, removed);
- Remove from `mPendingTransactions` // removeReceivedTxs(removed);

- Meanwhile, node keep listening for txs from other nodes as well and stores them into its `proposedSet` (could not verify it in code though)
- once the proposedSet is ready, its hash is computed  
- the computed hash, the next slot index and the proposedSet is added to the pendingEnvelopes
- After this NOMINATION starts // mNominationProtocol.nominate(value, previousValue, timedout);
- PREPARE -> COMMIT -> EXTERNALISE