Topic: NXT :: descendant of Bitcoin - Updated Information - page 1079. (Read 2761637 times)

And our second biggest bounty (after creating a finite mini-blockchain) should be to the team that is working hand in hand with the Cryptsy folks to create the software / data bridge necessary to get NXT on Cryptsy ASAP.  Who is this team?  Why have we not organized it?  Why are we not talking in this forum about their progress?

I don't think 1000 TPS is bandwidth limited, I think it is blockchain bloat limited.  As I've said before, 1000 TPS at 128 bytes per transaction is 10 GB added to the blockchain EVERY DAY if it is storing every transaction.  Anybody claim that math is wrong?  If that's correct, who is storing a complete copy of the blockchain?  It's not "service providers", since CfB has specifically said that "service providers" are not necessary for 1000 TPS.  I don't think it's going to be a Raspberry Pi or smartphone when its turn comes to forge a node - unless NXT blocks can be added to the NXT blockchain without a node having a complete local copy in memory?  Can somebody clarify to me whether or not this is true or possible?

On page one post one of this very thread we claim "Nxt improves different features of Bitcoin, removes some disadvantages (e.g. bloated blockchain)" and "Blockchain Shrinking - In progress".  I claim that if "moving from JSON to binary to achieve 128 byte transaction size on the blockchain" is our only bullet in this area, then NXT has big problems.

I had privately assumed (which in this case has made an ass of me and me alone) that NXT was incorporating JD Bruce's brilliant idea of a finite sized block chain based only on current account balances instead of historical transaction data.   Both CfB and serveral others including one I thought was BCNext were in on this idea at its inception, which from a timing standpoint was four to six months before NXT was announced.  

I thought finite mini-blockchain (along with transparent PoS forging) was the secret sauce in NXT.  Even JD himself takes note of NXT moving in this direction: https://bitcointalksearch.org/topic/m.3976934

Looks like I was wrong about that.

JD Bruce himself has a $20K bounty out to create the first finite mini-blockchain.  

March 2013 thread: Re: Blockchain-less P2P Currency (theoretical idea)
https://bitcointalk.org/index.php?topic=152662.40

April 2013 thread: Cryptocurrency with Finite "Mini-Blockchain"
https://bitcointalk.org/index.php?topic=169311.20

May 2013 Paper: Purely P2P Crypto-Currency With Finite Mini-Blockchain
http://www.bitfreak.info/files/pp2p-ccmbc-rev1.pdf

May 2013 thread: Re: Building the Next Generation of Crypto-Currency (developers required)
https://bitcointalksearch.org/topic/building-the-next-generation-of-crypto-currency-developers-required-215936

If BCNext and JLP and CfB have not already implemented a finite mini-blockchain behind the scenes, IMHO we need to issue our own NXT bounty for it, make it the biggest bounty we offer, and bump it to the top of our NXT add-on priority wish, WAY above adding yet another of these "services" we keep talking about.  

The blockchain is the fundamental foundation of NXT.  If we have not solved the blockchain bloat problem already (as we claim), we either do solve it in the future or we die.  I hope either somebody challenges me successfully on that point, or that we get our act together and start working on what's critical to our survival.   Who stores the complete 1000 TPS block chain?  Just how big can it get?  How do we prune this blockchain?  When and how often?  Can we make it finite sized and based on account balances and not transaction history?  

Please don't let me have brought this topic up yet again only to be ignored while the discussion veers off into something esoteric like which Turing complete instruction set to bolt onto NXT.

I want to have a nuts and bolts, meat and potatoes discussion about what infrastructure it's gonna take to support 1000 TPS.  I say we currently can't get there from here.  Somebody prove I'm full of crap.  Please.

As what i knew it was listed in cryptsy recently

you seem not to realize, how many people are actually waiting for some dumps

It happens very regularly in crypto.

Lands to Earth again?

What happens usually after a parabolic lift like that?

After TF is initialized there wouldn't really be a need to forge a block to complete a transaction, right? A privately signed IOU certificate could be passed along the network, verified by peers.

@ CfB

How much nxt would it cost to implement service provider feature before April 3rd?

GO ahead!!!

finally

CfB, I got question regarding assets.

I assume, network itself will do bid-ask matching.

I'm issuing an asset, user B is placing bid order.
Can "transferAsset" be used to "accept" the bid, without making an ask order?

If so, does it mean it should always include referencedTransaction?

I only have 800,000 NXT to dump and we r below 0,00005.

When I get back I will do this so i was polite warning you all.


oh and on vicrurex 90,000 NXT and we r below 0.00005 but that is to easy.

bye bye

For non commodities like say eBay listings,  then (2) should be acceptable.  However, when you do commodities or shares,  then the Bitcoin 10 minute per block limitation is really too restrictive. 

In general, you should do all the high speed stuff off the Bitcoin block chain and then resync back to the blockchain when done. 

See: http://www.nexcoin.net  on how it should really be done.

You probably want to take a look at NEX http://www.nexcoin.net .

It leverages the bitcoin secure network without the weight of Bitcoin's 10 minute confirmations.

I bought my first back on December 21 and yesterday was the first time I donated to a bounty (and I really only did it because someone donated some NXT to me for some quick work I did for them).

Having little NXT makes it frustrating when you see your small account balance versus all the worthy bounties and projects that need funding. :/ I just try to have fun, learn a little and watch the big dogs play in the yard as I watch from the porch.

Another issue of subleq, I can think of, is that it does not utilize hardware ressources efficiently.

Today's processors can calculate many different operations extremely efficiently and fast.

Our VM should utilize at least some of that potential.

Included in a block does not mean it's really confirmed. You still need to wait enough blocks to be sure the probability of reorganization is low.

To be fair:

1) The second point is valid, but you will not notice the difference as long as you have whole blockchain downloaded. When I pointed out this first in XCP thread, I mean the main drawback is the time to download the blockchain in the beginning. You will notice the slow parsing time only when a XCP is transferred tens of thousands of times, and that will not happen in near future.

2) The first point is of course true, but should be acceptable to most average users. Normally, we trade only once several days. Even day traders only trade couple of times per day.

We already sorted that out by: pay per op.