[0021] Propose a CKB address format

[CipherWang]

Thanks to github's 500 error, I have to recreate a new PR for CKB address format discussion.

[doitian]

Assign RFC# 21

[janx]

Please update number and category (should be in Standards Track).

RFC# assigned

[CipherWang]

Question: should we change the lock_hash format from blake256 to blake160 (leading 20 bytes of blake256)? It reduces address length dramatically without sacrifying security, since we use secp256k1 + blake160 as the default lock script.

Here is a length comparison between 256bit and 160bit lock hash for address.

256bit: ckb1qrxl9wt772fhrff5st8ew7kqcyceeshpqtn2eqv9a9eu3xvkkn40wv6tu2p
160bit: ckb1qrxl9wt772fhrff5st8ew7kqcyceeshpqgvgvdkq <= 19bytes shorter!

[CipherWang]

Please update number and category (should be in Standards Track).

updated.

updated.

[CipherWang]

1. remove lock_hash type address
2. add full lock script address support, and remove 90 characters limit of Bech32
3. compatible with previous address format

[janx]

• the file/directory names should be changed to reflect correct rfc number;
• please also add a row for this rfc in README

[CipherWang]

• the file/directory names should be changed to reflect correct rfc number;
• please also add a row for this rfc in README

done, please review again.

[ashchan]

🍻