[bitcoin-dev] Using a storage engine without UTXO-index

[Tomas]

On Tue, Apr 11, 2017, at 11:41, Eric Voskuil wrote:
> It's not the headers/tx-hashes of the blocks that I'm referring to, it
> is the confirmation and spend information relative to all txs and all
> outputs for each branch. This reverse navigation (i.e. utxo
> information) is essential, must be persistent and is branch-relative.

That is exactly what is stored in the spend-tree. 

>> As a simpler example, if two miners both mine a block at
>> approximately the same time and send it to each other, then surely
>> they would want to continue mining on their own block. Otherwise
>> they would be throwing away their own reward.

> That's not your concurrent validation scenario. In the scenario you
> described, the person chooses the weaker block of two that require
> validation because it's better somehow, not because it's his own
> (which does not require validation).

> Consistency is reached, despite seeing things at different times,
> because people use the same rules. If the economy ran on arbitrary
> block preference consistency would be elusive.

No but my example shows  that it is up to the miner to choose which tip
to work on. This is not using different rules, it is just optimizing its
income. This means that the economy *does* run on arbitrary "block
preference", even if it is not running on arbitrary rules.

If two blocks are competing, a miner could optimize its decision which
to mine on, not just on whether one of the blocks is his own, but also
on fees, or on excessive validation costs.

> I read this as encoding the height at which a fork historically
> activated. If you intend to track activation for each branch that will
> not be "height-based" it will be history based.

I understand "height-based" was not the right wording, as it is of
course branch-specific. Per tip ruleset metadata, must be matched with
per-transaction ruleset metadata.

Tomas