Quote from: kjj on May 30, 2013, 02:40:57 PM

Quote from: jtimon on May 30, 2013, 01:17:28 PM

If I'm understanding well, what he proposes is for the chain to maintain a ledger with pairs (address, total balance) and transactions just saying "substract X from account Y and add them to Z" instead of "spend everything in Y, X in Z and the rest back to Y". Very similar to Ripple's ledger.
This would be a huge refactor and a hardfork, but I haven't actually heard any argument or explanation why the system isn't like that from the beginning.
Saying "you don't know what we don't use GAAP because you don't know bitcoin" doesn't help.
Please, some one explain him (and me) why GAAP is worse than cash/change in any aspect.

Essentially, the ledger system offers no advantages (or only fictional advantages), and has a host of disadvantages. This really has been discussed to death. I'm not making that up to get rid of him, I just don't want to rehash the whole topic for the 99th time.

There are two main problems, and plenty others that are more subtle.

First, a ledger system provides no meaningful resources savings (no reduction in storage space, no reduction in memory usage, no reduction in network traffic, no reduction in CPU usage) because we are already very lean, meaning that our transactions are very close to the absolute minimum amount of information necessary.

Second, a ledger system is necessarily less flexible and useful. None of Mike's contracts would work in a ledger system without adding trusted third parties. You could regain that functionality by emulating the current scripting system with the scripts attached to disposable accounts rather than disposable transactions, but why bother? Bitcoin puts the complexity where it belongs, in the transaction.

As for accounting, there is nothing about bitcoin that is incompatible with it. The problem seems to come up when you try to map accounting functions to bitcoin functions based on their names. "Transaction" means something totally different in bitcoin than it does to an accountant. Same with "account", "balance", etc, etc. All of the ideas are compatible, as in, you can take an accounting-account and find an analogy in bitcoin, but that analog won't be a bitcoin-account, or a bitcoin-address, it will be something else, and that something else may not be "real" in bitcoin world.

I find the whole accounting discussion to be particularly tedious because the same thing is true in the real world, and yet, no one shows up on banking forums whining that bank-accounts are not exactly the same as accounting-accounts.

Quote

First, a ledger system provides no meaningful resources savings (no reduction in storage space, no reduction in memory usage, no reduction in network traffic, no reduction in CPU usage) because we are already very lean, meaning that our transactions are very close to the absolute minimum amount of information necessary.

See no reasoning for your daring statement, kjj.
As showed in detail using GAAP would save plenty of resources (~ 80%).
Estimated 50% just by avoiding these funny "changings", endless more using carry-overs if needed...

Quote

Second, a ledger system is necessarily less flexible and useful. None of Mike's contracts would work in a ledger system without adding trusted third parties. You could regain that functionality by emulating the current scripting system with the scripts attached to disposable accounts rather than disposable transactions, but why bother? Bitcoin puts the complexity where it belongs, in the transaction.

See nothing special in Mike Hearn's so-called "contracts", namely:
Providing a deposit, Escrow, Assurance contracts, Using external state, Trading across chains, Rapidly-adjusted (micro)payments to a pre-determined party or whatever else.

All quite normal commercial transactions. Not at all hindered or restrainted by GAAP. Mike doesn't claim that himself.
On the contrary!
Mikes examples show again how difficult it might be to solve normally not at all existing "problems" with current BTC.

https://en.bitcoin.it/wiki/Contracts