Hi Amir & list!

Thank you so much for having implemented libbitcoin, the library format
drives innovation.

Just wanted to ask you two things:

(A) How stable is libbitcoin?:

How much has it been used, any particular site or project that you felt
has given it a particular 'fire proof'?;

With what care with regard to security was it implemented?;

Would you say that it is mature to reliably run a production fully
verifying node on it?;

Do you intend to maintain it to be spinning as it should all the time
(considering any security/fork/other bugs) over the years to come?

Do you feel any relevancy whatsoever currently in running it behind a
bitcoind for firewall and easier blockchain restore?

(B) About how would you suggest using libbitcoin or/and Obelisk for

* Lots of addrindex lookup for sum of unspent BTC on given BTC
addresses

So like, a heap-based array of like 1-100k BTC addresses, each with
unspent sum.

* Counting confirmations (preferably even on a
per-remote-node-connected-to basis) on Lots of given transactions?

So like, a heap-based array of like 1k TX ID:s, each with confirmation
count. It doesn't need to count to more than 100 or so. Should be pretty
fast for any transaction.

Which of libbtc or/and Obelisk would you use, about which
classes/interface, about how would you structure this?

(any code/thread/process design is fine)

pointers to docs/examples would be much appreciated.

Thanks!