embedding the hash of UTXO in the block headers and retaining only the last few hundred blocks is being experimented with as a client option.
Good to know.
screw that, i want direct access to a copy of the entire blockchain at all times.
No one will deny you that access on an open network should you wish it ... destruction of tx history has a myriad of benefits, and risks (that can be mitigated).
Yes. The power of Bitcoin is enhanced by in-built flexibility. Many nodes will always have and always want the full blockchain, many will be happier with the UTXO set secured by hashes in the main-chain, and many users will be OK with SPV only.
Just because alternative models are available, and can work concurrently with earlier implementations, does not mean everyone jumps ship to one model.This is exactly why I like UTXO hashes embedded in blocks. It does not preclude nodes from performing full chain scans, it simply extends and provides additional mechanisms to participate in the P2P network.
With UTXO hashes you could think of bitcoin extending to 3 types of nodes, where today we only have two. They would be:
1) SPV clients - Light-clients, do not protect/valid network, protect own private keys only
2) Full nodes - Full participants on P2P network,
protect and valid the current ledger only (which is all that is needed to be a node)
3) Archival nodes - Full participants on P2P network,
protect and valid the full historical journal (as Marcus called it)
UTXO hashes would enable many more people to participate as P2P nodes and actively add protection to the current ledger. This does not stop several archival nodes from existing who would act as maintainers of the historical journal for prosperity and validation.
By creating this division more people/nodes can participate which improves bitcoin's security.
