A standard transaction is 225 bytes, leading to a savings of 8.6%.

However, that is essentially the minimum saving. For other sizes (eg, 10 outputs) which seem to be pretty frequent savings can be greater.

Furthermore, it is important to note that this is a memory saving for the UTXO pool so the reduction there is more (not super familiar with how the UTXO pool is stored, but it should be a better savings than 8.6%). 


Does anyone have the tools currently to be able to easily run through the chain and analyze the impact of this change on historical data more directly?

On Thu, Jul 23, 2015 at 12:55 PM, jl2012 via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:

Quoting Gavin Andresen via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org>:

On Wed, Jul 22, 2015 at 4:34 PM, Tier Nolan via bitcoin-dev <
bitcoin-dev@lists.linuxfoundation.org> wrote:

It also requires most clients to be updated to support the new address
system.


That's the killer: introducing Yet Another Type of Bitcoin Address takes a
very long time and requires a lot of people to change their code. At least,
that was the lesson learned when we introduced P2SH addresses.

I think it's just not worth it for a very modest space savings (10 bytes,
when scriptSig+scriptPubKey is about 120 bytes), especially with the
extreme decrease in security (going from 2^160 to 2^80 to brute-force).

--
--
Gavin Andresen

I think it would only save ~5% with all overhead (value, sequence, locktime, version, etc.) counted

A better way is to introduce shorter ECDSA keys, which will save a lot of space for public key and signature. It is safe as long as the output value is much lower than the cost of attack.

If this happens, I think it will be part of the OP_MAST which will require a new address type anyway.


_______________________________________________
bitcoin-dev mailing list
bitcoin-dev@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev