leveldb would not improve the start-up time by much, as there are some more PoS-related checks and also the scrypt Nfactor a major role in load times (and with so many blocks it gets even worse).
yeah, the geolocation data was on my old site - though afaik it was quite inaccurate (a LONG delay before an inactive address was removed from the list - so it was a bit inflated). i could make a more accurate version of that from the dnsseed server data (currently ~400 nodes, just 35 with over 10% uptime over last 30 days - so the network isn't that big).
no opinion on the qt wallet transparency here
aliases - i'd rather not do that. even firstbits seemed like a dumb idea to me. i can guarantee it'll end up like the coinbase's aliases (at least i think it was coinbase) - a fake sean's outpost charity alias registered within minutes and "the ceo of bitcoin", to name a few
what i'd like to see is people swithing from full qt wallet to thin clients (instant start-up, negligible storage requirements, not so much cpu-intensive - so it can run on your smartphone without draining the battery in seconds). however, that would require a PoS minting capability built into such clients - which is not an easy task, but certainly doable. just my 2 yacs
imo the node "buckets" and overall connection policy in bitcoin and derivatives in quite good and i don't see anything obvious to improve there. traffic analysis? there's no encryption whatsoever, so i don't see the point there. same with sigint. also, all the *coin networks have a unique identification "magic" so you can eg. easily block all bitcoin traffic with a simple deep packet inspection.
yeah, the geolocation data was on my old site - though afaik it was quite inaccurate (a LONG delay before an inactive address was removed from the list - so it was a bit inflated). i could make a more accurate version of that from the dnsseed server data (currently ~400 nodes, just 35 with over 10% uptime over last 30 days - so the network isn't that big).
no opinion on the qt wallet transparency here
aliases - i'd rather not do that. even firstbits seemed like a dumb idea to me. i can guarantee it'll end up like the coinbase's aliases (at least i think it was coinbase) - a fake sean's outpost charity alias registered within minutes and "the ceo of bitcoin", to name a few
what i'd like to see is people swithing from full qt wallet to thin clients (instant start-up, negligible storage requirements, not so much cpu-intensive - so it can run on your smartphone without draining the battery in seconds). however, that would require a PoS minting capability built into such clients - which is not an easy task, but certainly doable. just my 2 yacs
With many full nodes that accept incoming connections some cool "anon" features can be done. Not a single altcoin yet improved a way nodes connect and reconnect over time which is what makes networks an easy target for signal inteligence and especially traffic analysis.
imo the node "buckets" and overall connection policy in bitcoin and derivatives in quite good and i don't see anything obvious to improve there. traffic analysis? there's no encryption whatsoever, so i don't see the point there. same with sigint. also, all the *coin networks have a unique identification "magic" so you can eg. easily block all bitcoin traffic with a simple deep packet inspection.