mrkavasaki
|
|
July 31, 2015, 03:01:50 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin
|
|
|
|
Panadacoin
|
|
July 31, 2015, 03:06:06 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin What is the point to come and spread fud here?
|
|
|
|
kadrek
|
|
July 31, 2015, 03:10:03 AM |
|
What is the point to come and spread fud here?
They feel their coins are threatened by Dash and are envious of everyone who got in early
|
SWIPE │ │ │ Monetizing mobile engagement data, on the blockchain [ SWIPE . WHITEPAPER ] TELEGRAM TWITTER MEDIUM REDDIT
|
|
|
eduffield (OP)
Legendary
Offline
Activity: 1176
Merit: 1036
Dash Developer
|
|
July 31, 2015, 03:12:42 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin Once the budget system is running we're going to have bounties available. How about $2000 to break an 8 round DS? It's never been done because it's impossible
|
Dash - Digital Cash | dash.org | dashfoundation.io | dashgo.io
|
|
|
TheDasher
|
|
July 31, 2015, 03:19:50 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin Once the budget system is running we're going to have bounties available. How about $2000 to break an 8 round DS? It's never been done because it's impossible Hardly worth anyone's time at 2000. If you are certain of your product why not up the ante to something which might motivate someone with half a brain to try?
|
|
|
|
TanteStefana2
Legendary
Offline
Activity: 1260
Merit: 1001
|
|
July 31, 2015, 04:49:10 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin Once the budget system is running we're going to have bounties available. How about $2000 to break an 8 round DS? It's never been done because it's impossible Sure, but lets not put too much into bounties that may never be claimed 'cause I'd rather see the coins get used and distributed. But I'm kind of hammering the head off the nail with this one, LOL.
|
Another proud lifetime Dash Foundation member My TanteStefana account was hacked, Beware trading "You'll never reach your destination if you stop to throw stones at every dog that barks."Sir Winston Churchill BTC: 12pu5nMDPEyUGu3HTbnUB5zY5RG65EQE5d
|
|
|
sdef2
Member
Offline
Activity: 60
Merit: 10
|
|
July 31, 2015, 06:12:01 AM |
|
looks great! By the way, at 2:05 in the video we can see you even had a 0.1 second 'sleep' command after each send, i was wondering why ?
|
|
|
|
Macrochip
|
|
July 31, 2015, 06:40:14 AM |
|
This was an almost turd-free zone for days. All Evan had to do was tear the haters a new one with a tiny video clip and the turds came zooming out like torpedoes leaving their skidmarks with age old debunked and refuted shitstains disguised as arguments. Marvelous how the hate and envy machine of Failero works... Eat your heart out, losers. Why would eduffield not relaunch Because the community voted "NO" you lying criminal scumbag. Go to jail and give the money back you stole, Hashfast scammer!dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin Oh hey, nice Trust level you got there, bro! Why is it always liars, thiefs and scammers who attack Dash?
|
|
|
|
illodin
|
|
July 31, 2015, 07:09:42 AM |
|
dash is not even really anonymous, it will replaced sooner or later by an cryptonode coin Once the budget system is running we're going to have bounties available. How about $2000 to break an 8 round DS? It's never been done because it's impossible Hardly worth anyone's time at 2000. If you are certain of your product why not up the ante to something which might motivate someone with half a brain to try? Are you saying there's no one with half a brain in Monero who holds enough Moneros to have the upside of +$2000 when the big bad DASH goes away so MONERO can go to MOOOON?
|
|
|
|
illodin
|
|
July 31, 2015, 07:11:26 AM |
|
Greed addresses and InstantX are totally different. For starters, Green addresses require TRUST, and each trusted party have to be trusted separately. Very unpractical.
Whereas InstantX is trustless, and just works automatically for everyone.
|
|
|
|
Sub-Ether
Sr. Member
Offline
Activity: 434
Merit: 250
Quantum entangled and jump drive assisted messages
|
|
July 31, 2015, 07:35:13 AM |
|
Looks infinitely scalable to me, might have to test 1 million instantX transactions just for the sheer hell of it ' So how many tx a quorum(10 MNs) can take? Is it a matter of bandwidth or...please axplain?? If 30000 or more itx hit MN network, what then? Can a MN on average be a part of 3000 quorums or more?' --Lukas_Jackson-- ' Sure, the only limit to the system is the amount of memory we have available. But then again, we have a 2-tier network... so we can just require beefy masternodes. ' --Evan Duffield--
|
Dash is 27.3 times faster with syncing and updating than Bitcoin and 93.7 times faster than Monero. Bitcoin (v0.11.0) has a Tao ratio 11.2% faster than bitcoin (v0.10.0) release. Dash (v.0.12.0.49) = Tao sync ratio = 0.15 seconds / hour of update || Dash (v.0.11.2.23) = Tao sync ratio = 0.24 seconds / hour of update. V12 versus V11 speedup = +36.5% Bitcoin (v.0.11.0) = Tao sync ratio = 4.14 seconds / hour of update || Monero (v.0.41.1) = Tao sync ratio = 14.2 seconds / hour of update
|
|
|
thelonecrouton
Legendary
Offline
Activity: 966
Merit: 1000
|
|
July 31, 2015, 07:41:46 AM |
|
Bitcoin currently handles 7 unconfirmed tx per second + confirmation/waiting time DASH handles 10 confirmed tx per second and you can walk out the shop with the merchant safe and happy.
Not quite right, the entire Bitcoin network handles a maximum of 7 txes per second. Each IX only requires a tiny subset of the 3000 MN total to work. You're looking at tx throughput two orders of magnitude+ better than BTC, each one in seconds, not minutes/hours.
|
|
|
|
thelonecrouton
Legendary
Offline
Activity: 966
Merit: 1000
|
|
July 31, 2015, 08:00:41 AM |
|
Why does the network have so much new capacity with this release ?
IX capacity has always been better than PoW capacity, as you're not having to wait for the whole lumbering Proof-of-Waste farce, you're getting consensus from a much smaller group of nodes, with the main constraint being network latency (milliseconds), not arbitrary Watt-burning make-work (minutes+.)
|
|
|
|
Macrochip
|
|
July 31, 2015, 08:01:01 AM |
|
Bitcoin currently handles 7 unconfirmed tx per second + confirmation/waiting time DASH handles 10 confirmed tx per second and you can walk out the shop with the merchant safe and happy.
Not quite right, the entire Bitcoin network handles a maximum of 7 txes per second. Each IX only requires a tiny subset of the 3000 MN total to work. You're looking at tx throughput two orders of magnitude+ better than BTC, each one in seconds, not minutes/hours. So if the subset is 10 for example, at 3k M-Nodes the TPS would be 3,000 as well, I assume? That's 1k more than VISA handles atm. This is a very interesting read on that matter: https://en.bitcoin.it/wiki/Scalability
|
|
|
|
thelonecrouton
Legendary
Offline
Activity: 966
Merit: 1000
|
|
July 31, 2015, 08:06:34 AM |
|
Bitcoin currently handles 7 unconfirmed tx per second + confirmation/waiting time DASH handles 10 confirmed tx per second and you can walk out the shop with the merchant safe and happy.
Not quite right, the entire Bitcoin network handles a maximum of 7 txes per second. Each IX only requires a tiny subset of the 3000 MN total to work. You're looking at tx throughput two orders of magnitude+ better than BTC, each one in seconds, not minutes/hours. So if the subset is 10 for example, at 3k M-Nodes the TPS would be 3,000 as well, I assume? That's 1k more than VISA handles atm. This is a very interesting read on that matter: https://en.bitcoin.it/wiki/ScalabilityIf 10 MNs are needed for each IX then with 3000 MNs you'd have theoretical capacity for 300 simultaneous IX's. There are going to be overheads, and some MNs may be tied up doing DS or something else, but you get the idea. I don't know if or how the concept of MN multithreading/multitasking applies, but there's probably room to squeeze a shitload more efficiency out of the system yet.
|
|
|
|
Macrochip
|
|
July 31, 2015, 08:09:53 AM |
|
Why does the network have so much new capacity with this release ?
IX capacity has always been better than PoW capacity, as you're not having to wait for the whole lumbering Proof-of-Waste farce, you're getting consensus from a much smaller group of nodes. Without PoW consensus I feel like this is becoming dangerous territory: If the subset of elected IX-nodes is small enough, wouldn't someone like Otoh with over 600 MNs have a good chance that a TX is entirely in his domain? What would stop an attacker from reversing an IX-tx before it hits the miners, when all consensus nodes are controlled by one person?
|
|
|
|
thelonecrouton
Legendary
Offline
Activity: 966
Merit: 1000
|
|
July 31, 2015, 08:14:01 AM Last edit: July 31, 2015, 08:27:46 AM by thelonecrouton |
|
Why does the network have so much new capacity with this release ?
IX capacity has always been better than PoW capacity, as you're not having to wait for the whole lumbering Proof-of-Waste farce, you're getting consensus from a much smaller group of nodes. Without PoW consensus I feel like this is becoming dangerous territory: If the subset of elected IX-nodes is small enough, wouldn't someone like Otoh with over 600 MNs have a good chance that a TX is entirely in his domain? What would stop an attacker from reversing an IX-tx before it hits the miners, when all consensus nodes are controlled by one person? You have it exactly backwards, MN subsetting is tens of thousands of times more secure than PoW. If there are 600 compromised nodes out of 3000, and you need 10 of those 3000 to beat the system, the chances that all of the 10 needed nodes for any one IX will be among the 600 compromised nodes are miniscule. Roughly, (600/3000)^10 = 0.000000102 Evil-Otoh will be waiting a long time to bugger up a single IX. There's a reason hierarchical structures arise in nature and society - efficiency. Societal hierarchies are easy targets for compromise because they are essentially static, they provide fixed targets. Pooled mining is a good example of an this, it's a vulnerability, not an asset. Overlay networks like Masternodes are a fluid hierarchy and vastly more resilient to attack.
|
|
|
|
Sub-Ether
Sr. Member
Offline
Activity: 434
Merit: 250
Quantum entangled and jump drive assisted messages
|
|
July 31, 2015, 08:24:30 AM |
|
Bitcoin currently handles 7 unconfirmed tx per second + confirmation/waiting time DASH handles 10 confirmed tx per second and you can walk out the shop with the merchant safe and happy.
Not quite right, the entire Bitcoin network handles a maximum of 7 txes per second. Each IX only requires a tiny subset of the 3000 MN total to work. You're looking at tx throughput two orders of magnitude+ better than BTC, each one in seconds, not minutes/hours. So if the subset is 10 for example, at 3k M-Nodes the TPS would be 3,000 as well, I assume? That's 1k more than VISA handles atm. This is a very interesting read on that matter: https://en.bitcoin.it/wiki/ScalabilityIf 10 MNs are needed for each IX then with 3000 MNs you'd have theoretical capacity for 300 simultaneous IX's. There are going to be overheads, and some MNs may be tied up doing DS or something else, but you get the idea. I don't know if or how the concept of MN multithreading/multitasking applies, but there's probably room to squeeze a shitload more efficiency out of the system yet. 300 is not the maximum, Evan said its down to the amount of memory of the masternodes, the upper limit is unknown. We were sending hundreds of IX in testnet with only 150 masternodes and they all worked up until someone reduced the number of nodes, there looked to be a failure point were they return to proof of work but its pretty low node count wise and when scaled up to 3000 nodes may be impossible to break due to the strength of the network.
|
Dash is 27.3 times faster with syncing and updating than Bitcoin and 93.7 times faster than Monero. Bitcoin (v0.11.0) has a Tao ratio 11.2% faster than bitcoin (v0.10.0) release. Dash (v.0.12.0.49) = Tao sync ratio = 0.15 seconds / hour of update || Dash (v.0.11.2.23) = Tao sync ratio = 0.24 seconds / hour of update. V12 versus V11 speedup = +36.5% Bitcoin (v.0.11.0) = Tao sync ratio = 4.14 seconds / hour of update || Monero (v.0.41.1) = Tao sync ratio = 14.2 seconds / hour of update
|
|
|
thelonecrouton
Legendary
Offline
Activity: 966
Merit: 1000
|
|
July 31, 2015, 08:32:14 AM |
|
Bitcoin currently handles 7 unconfirmed tx per second + confirmation/waiting time DASH handles 10 confirmed tx per second and you can walk out the shop with the merchant safe and happy.
Not quite right, the entire Bitcoin network handles a maximum of 7 txes per second. Each IX only requires a tiny subset of the 3000 MN total to work. You're looking at tx throughput two orders of magnitude+ better than BTC, each one in seconds, not minutes/hours. So if the subset is 10 for example, at 3k M-Nodes the TPS would be 3,000 as well, I assume? That's 1k more than VISA handles atm. This is a very interesting read on that matter: https://en.bitcoin.it/wiki/ScalabilityIf 10 MNs are needed for each IX then with 3000 MNs you'd have theoretical capacity for 300 simultaneous IX's. There are going to be overheads, and some MNs may be tied up doing DS or something else, but you get the idea. I don't know if or how the concept of MN multithreading/multitasking applies, but there's probably room to squeeze a shitload more efficiency out of the system yet. 300 is not the maximum, Evan said its down to the amount of memory of the masternodes, the upper limit is unknown. We were sending hundreds of IX in testnet with only 150 masternodes and they all worked up until someone reduced the number of nodes, there looked to be a failure point were they return to proof of work but its pretty low node count wise and when scaled up to 3000 nodes may be impossible to break due to the strength of the network. Splendid. Three+ orders of magnitude better tx throughput than Bitcoin then.
|
|
|
|
Macrochip
|
|
July 31, 2015, 08:48:04 AM |
|
If there are 600 compromised nodes out of 3000, and you need 10 of those 3000 to beat the system, the chances that all of the 10 needed nodes for any one IX will be among the 600 compromised nodes are miniscule.
Roughly, (600/3000)^10 = 0.000000102
Evil-Otoh will be waiting a long time to bugger up a single IX.
Just covering all bases here. Thanks for the explanation. I really have to refresh my combinatoric knowledge
|
|
|
|
|