Yes, absolutely, good clarification.

The protocol has it covered and there is room for additional
optimization by design through extension of the transaction return
enumeration.

e

On 09/25/2014 09:43 PM, William Swanson wrote:
> In a sense, we already have "get_history" in the new protocol - just
> do a "get transactions" query starting from block 0 with no result
> limit, and specifying an address "prefix" that's 160 bits long (in
> other words, the whole address). This will return all transactions for
> a particular address. The client can trivially match inputs to outputs
> and build the `history_list` structure we have today.
>
> Of course, that's a bit more expensive from a bandwidth point of view
> than what we have now. It may be that we add a new "summary" option to
> the result_type structure as an optimization, as Eric suggests. Or,
> maybe we remove the "utxo" result type and replace it with "summary."
> We'll have to give this protocol some real-world use before we can
> come to any sort of definite conclusion.
>
> -William
>
> On Wed, Sep 24, 2014 at 6:34 PM, Eric Voskuil <eric@???> wrote:
>> I assume that by fixed you mean full/not-prefixed addresses. Based on
>> that assumption I see it as a two part question:
>>
>> 1. non-private query by address
>>
>> In development of the new protocol we agreed that as a rule there should
>> be no queries that must be used in a non-private manner. IOW all queries
>> would be "prefixed" and only if a caller choses to trade privacy for
>> performance would all bits of the prefix be specified.
>>
>> 2. query for summary (history) information
>>
>> A query for summary information is useful for a stateless server
>> trusting client, but not for full-chain, spv or caching server trusting
>> clients.
>>
>> https://github.com/libbitcoin/libbitcoin_protocol#client-type
>>
>> We incorporated such support into the protocol, although only in the
>> form of utxo results, not complete history. We should probably add a
>> history_result to the results enumeration.
>>
>> As we are working on SPV and require private queries I would expect to
>> use this in our wallet.
>>
>> e
>>
>> On 09/24/2014 06:07 PM, Amir Taaki wrote:
>>> changed it over for the new spv-api.
>>>
>>> What do you and William think about a get_history command for fixed
>>> addresses?
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Libbitcoin mailing list
>>> Libbitcoin@???
>>> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/libbitcoin
>>>
>>
>>
>> _______________________________________________
>> Libbitcoin mailing list
>> Libbitcoin@???
>> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/libbitcoin
>>